feat: support gemini ai model (#1173 )

fix: fix tracing configmap template to handle initial installation (#1191 )
feat: improve Skywalking and Zipkin integration (#1131 )
2026-03-08 02:30:56 +08:00 · 2024-08-09 09:55:40 +08:00 · 2024-08-09 08:29:51 +08:00 · 2024-08-08 22:40:33 +08:00 · 2024-08-08 18:04:47 +08:00 · 2024-08-08 18:04:33 +08:00
279 changed files with 15188 additions and 788 deletions
--- a/.github/workflows/build-and-push-wasm-plugin-image.yaml
+++ b/.github/workflows/build-and-push-wasm-plugin-image.yaml
@@ -0,0 +1,114 @@
+name: Build and Push Wasm Plugin Image
+
+on:
+  push:
+    tags:
+    - "wasm-go-*-v*.*.*" # 匹配 wasm-go-{pluginName}-vX.Y.Z 格式的标签
+  workflow_dispatch:
+    inputs:
+      plugin_name:
+        description: 'Name of the plugin'
+        required: true
+        type: string
+      version:
+        description: 'Version of the plugin (optional, without leading v)'
+        required: false
+        type: string
+
+jobs:
+  build-and-push-wasm-plugin-image:
+    runs-on: ubuntu-latest
+    environment:
+      name: image-registry-msg
+    env:
+      IMAGE_REGISTRY_SERVICE: ${{ vars.IMAGE_REGISTRY || 'higress-registry.cn-hangzhou.cr.aliyuncs.com' }}
+      IMAGE_REPOSITORY: ${{ vars.PLUGIN_IMAGE_REPOSITORY || 'plugins' }}
+      GO_VERSION: 1.19
+      TINYGO_VERSION: 0.28.1
+      ORAS_VERSION: 1.0.0
+    steps:
+      - name: Set plugin_name and version from inputs or ref_name
+        id: set_vars
+        run: |
+          if [[ "${{ github.event_name }}" == "workflow_dispatch" ]]; then
+            plugin_name="${{ github.event.inputs.plugin_name }}"
+            version="${{ github.event.inputs.version }}"
+          else
+            ref_name=${{ github.ref_name }}
+            plugin_name=${ref_name#*-*-} # 删除插件名前面的字段(wasm-go-)
+            plugin_name=${plugin_name%-*} # 删除插件名后面的字段(-vX.Y.Z)
+            version=$(echo "$ref_name" | awk -F'v' '{print $2}')
+          fi
+
+          echo "PLUGIN_NAME=$plugin_name" >> $GITHUB_ENV
+          echo "VERSION=$version" >> $GITHUB_ENV
+
+      - name: Checkout code
+        uses: actions/checkout@v3
+      
+      - name: File Check
+        run: | 
+          workspace=${{ github.workspace }}/plugins/wasm-go/extensions/${PLUGIN_NAME}
+          push_command="./plugin.tar.gz:application/vnd.oci.image.layer.v1.tar+gzip"
+
+          # 查找spec.yaml
+          if [ -f "${workspace}/spec.yaml" ]; then
+            echo "spec.yaml exists"
+            push_command="./spec.yaml:application/vnd.module.wasm.spec.v1+yaml $push_command "
+          fi
+
+          # 查找README.md
+          if [ -f "${workspace}/README.md" ];then
+              echo "README.md exists"
+              push_command="./README.md:application/vnd.module.wasm.doc.v1+markdown $push_command "
+          fi
+          
+          # 查找README_{lang}.md
+          for file in ${workspace}/README_*.md; do
+            if [ -f "$file" ]; then
+              file_name=$(basename $file)
+              echo "$file_name exists"
+              lang=$(basename $file | sed 's/README_//; s/.md//')
+              push_command="./$file_name:application/vnd.module.wasm.doc.v1.$lang+markdown $push_command "
+            fi
+          done
+
+          echo "PUSH_COMMAND=\"$push_command\"" >> $GITHUB_ENV
+        
+      - name: Run a wasm-go-builder
+        env: 
+          PLUGIN_NAME: ${{ env.PLUGIN_NAME }}
+          BUILDER_IMAGE: higress-registry.cn-hangzhou.cr.aliyuncs.com/plugins/wasm-go-builder:go${{ env.GO_VERSION }}-tinygo${{ env.TINYGO_VERSION }}-oras${{ env.ORAS_VERSION }}
+        run: |
+          docker run -itd --name builder -v ${{ github.workspace }}:/workspace -e PLUGIN_NAME=${{ env.PLUGIN_NAME }} --rm ${{ env.BUILDER_IMAGE }} /bin/bash
+
+      - name: Build Image and Push
+        run: |
+          push_command=${{ env.PUSH_COMMAND }}
+          push_command=${push_command#\"}
+          push_command=${push_command%\"} # 删除PUSH_COMMAND中的双引号，确保oras push正常解析
+          
+          target_image="${{ env.IMAGE_REGISTRY_SERVICE }}/${{ env.IMAGE_REPOSITORY}}/${{ env.PLUGIN_NAME }}:${{ env.VERSION }}"
+          echo "TargetImage=${target_image}"
+
+          cd ${{ github.workspace }}/plugins/wasm-go/extensions/${PLUGIN_NAME}
+          if [ -f ./.buildrc ]; then
+            echo 'Found .buildrc file, sourcing it...'
+            . ./.buildrc
+          else
+            echo '.buildrc file not found'
+          fi
+          echo "EXTRA_TAGS=${EXTRA_TAGS}"
+
+          command="
+          set -e
+          cd /workspace/plugins/wasm-go/extensions/${PLUGIN_NAME}
+          go mod tidy
+          tinygo build -o ./plugin.wasm -scheduler=none -target=wasi -gc=custom -tags=\"custommalloc nottinygc_finalizer ${EXTRA_TAGS}\" .
+          tar czvf plugin.tar.gz plugin.wasm
+          echo ${{ secrets.REGISTRY_PASSWORD }} | oras login -u ${{ secrets.REGISTRY_USERNAME }} --password-stdin ${{ env.IMAGE_REGISTRY_SERVICE }}
+          oras push ${target_image} ${push_command}
+          "
+          docker exec builder bash -c "$command"
+
+          
--- a/.github/workflows/build-and-test-plugin.yaml
+++ b/.github/workflows/build-and-test-plugin.yaml
@@ -17,8 +17,8 @@ jobs:
  lint:
    runs-on: ubuntu-latest
    steps:
-      - uses: actions/checkout@v3
-      - uses: actions/setup-go@v3
+      - uses: actions/checkout@v4
+      - uses: actions/setup-go@v5
        with:
          go-version: 1.19
    # There are too many lint errors in current code bases
@@ -30,9 +30,9 @@ jobs:
    strategy:
      matrix:
        # TODO(Xunzhuo): Enable C WASM Filters in CI
-        wasmPluginType: [ GO ]
+        wasmPluginType: [ GO, RUST ]
    steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4

      - name: Free Up GitHub Actions Ubuntu Runner Disk Space 🔧
        uses: jlumbroso/free-disk-space@main
@@ -45,12 +45,12 @@ jobs:
          swap-storage: true      

      - name: "Setup Go"
-        uses: actions/setup-go@v3
+        uses: actions/setup-go@v5
        with:
          go-version: 1.19

      - name: Setup Golang Caches
-        uses: actions/cache@v3
+        uses: actions/cache@v4
        with:
          path: |-
            ~/.cache/go-build
@@ -60,7 +60,7 @@ jobs:
            ${{ runner.os }}-go

      - name: Setup Submodule Caches
-        uses: actions/cache@v3
+        uses: actions/cache@v4
        with:
          path: |-
            .git/modules
@@ -81,4 +81,4 @@ jobs:
    runs-on: ubuntu-latest
    needs: [ higress-wasmplugin-test ]
    steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
--- a/.github/workflows/build-and-test.yaml
+++ b/.github/workflows/build-and-test.yaml
@@ -10,8 +10,8 @@ jobs:
  lint:
    runs-on: ubuntu-latest
    steps:
-    - uses: actions/checkout@v3
-    - uses: actions/setup-go@v3
+    - uses: actions/checkout@v4
+    - uses: actions/setup-go@v5
      with:
        go-version: 1.19
    # There are too many lint errors in current code bases
@@ -21,10 +21,10 @@ jobs:
  coverage-test:
    runs-on: ubuntu-latest
    steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4

    - name: Setup Golang Caches
-      uses: actions/cache@v3
+      uses: actions/cache@v4
      with:
        path: |-
          ~/.cache/go-build
@@ -33,7 +33,7 @@ jobs:
        restore-keys: ${{ runner.os }}-go

    - name: Setup Submodule Caches
-      uses: actions/cache@v3
+      uses: actions/cache@v4
      with:
        path: |-
            .git/modules
@@ -46,7 +46,7 @@ jobs:
    - name: Run Coverage Tests
      run: GOPROXY="https://proxy.golang.org,direct" make go.test.coverage
    - name: Upload coverage to Codecov
-      uses: codecov/codecov-action@v3
+      uses: codecov/codecov-action@v4
      with:
        fail_ci_if_error: false
        files: ./coverage.xml
@@ -58,17 +58,17 @@ jobs:
    needs: [lint,coverage-test]
    steps:
      - name: "Checkout ${{ github.ref }}"
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
        with:
          fetch-depth: 2

      - name: "Setup Go"
-        uses: actions/setup-go@v3
+        uses: actions/setup-go@v5
        with:
          go-version: 1.19

      - name: Setup Golang Caches
-        uses: actions/cache@v3
+        uses: actions/cache@v4
        with:
          path: |-
            ~/.cache/go-build
@@ -77,7 +77,7 @@ jobs:
          restore-keys: ${{ runner.os }}-go

      - name: Setup Submodule Caches
-        uses: actions/cache@v3
+        uses: actions/cache@v4
        with:
          path: |-
            .git/modules
@@ -90,7 +90,7 @@ jobs:
        run: GOPROXY="https://proxy.golang.org,direct" make build

      - name: Upload Higress Binary
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
        with:
          name: higress
          path: out/
@@ -108,12 +108,12 @@ jobs:
    - uses: actions/checkout@v3
      
    - name: "Setup Go"
-      uses: actions/setup-go@v3
+      uses: actions/setup-go@v5
      with:
        go-version: 1.19

    - name: Setup Golang Caches
-      uses: actions/cache@v3
+      uses: actions/cache@v4
      with:
        path: |-
          ~/.cache/go-build
@@ -123,7 +123,7 @@ jobs:
          ${{ runner.os }}-go
      
    - name: Setup Submodule Caches
-      uses: actions/cache@v3
+      uses: actions/cache@v4
      with:
        path: |-
            .git/modules
@@ -139,4 +139,4 @@ jobs:
    runs-on: ubuntu-latest
    needs: [higress-conformance-test,gateway-conformance-test]
    steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
--- a/.github/workflows/build-image-and-push.yaml
+++ b/.github/workflows/build-image-and-push.yaml
@@ -16,7 +16,7 @@ jobs:
      CONTROLLER_IMAGE_NAME: ${{ vars.CONTROLLER_IMAGE_NAME || 'higress/higress' }}
    steps:
      - name: "Checkout ${{ github.ref }}"
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
        with:
          fetch-depth: 1

@@ -31,12 +31,12 @@ jobs:
          swap-storage: true

      - name: "Setup Go"
-        uses: actions/setup-go@v3
+        uses: actions/setup-go@v5
        with:
          go-version: 1.19

      - name: Setup Golang Caches
-        uses: actions/cache@v3
+        uses: actions/cache@v4
        with:
          path: |-
            ~/.cache/go-build
@@ -45,7 +45,7 @@ jobs:
          restore-keys: ${{ runner.os }}-go

      - name: Setup Submodule Caches
-        uses: actions/cache@v3
+        uses: actions/cache@v4
        with:
          path: |-
            envoy
@@ -56,7 +56,7 @@ jobs:

      - name: Calculate Docker metadata
        id: docker-meta
-        uses: docker/metadata-action@v4
+        uses: docker/metadata-action@v5
        with:
          images: |
            ${{ env.CONTROLLER_IMAGE_REGISTRY }}/${{ env.CONTROLLER_IMAGE_NAME }}
@@ -67,7 +67,7 @@ jobs:
            type=raw,value=latest,enable=${{ github.ref == format('refs/heads/{0}', 'main') }}

      - name: Login to Docker Registry
-        uses: docker/login-action@v2
+        uses: docker/login-action@v3
        with:
          registry: ${{ env.CONTROLLER_IMAGE_REGISTRY }}
          username: ${{ secrets.REGISTRY_USERNAME }}
@@ -92,7 +92,7 @@ jobs:
      PILOT_IMAGE_NAME: ${{ vars.PILOT_IMAGE_NAME || 'higress/pilot' }}
    steps:
      - name: "Checkout ${{ github.ref }}"
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
        with:
          fetch-depth: 1

@@ -107,12 +107,12 @@ jobs:
          swap-storage: true

      - name: "Setup Go"
-        uses: actions/setup-go@v3
+        uses: actions/setup-go@v5
        with:
          go-version: 1.19

      - name: Setup Golang Caches
-        uses: actions/cache@v3
+        uses: actions/cache@v4
        with:
          path: |-
            ~/.cache/go-build
@@ -121,7 +121,7 @@ jobs:
          restore-keys: ${{ runner.os }}-go

      - name: Setup Submodule Caches
-        uses: actions/cache@v3
+        uses: actions/cache@v4
        with:
          path: |-
            envoy
@@ -132,7 +132,7 @@ jobs:

      - name: Calculate Docker metadata
        id: docker-meta
-        uses: docker/metadata-action@v4
+        uses: docker/metadata-action@v5
        with:
          images: |
            ${{ env.PILOT_IMAGE_REGISTRY }}/${{ env.PILOT_IMAGE_NAME }}
@@ -143,7 +143,7 @@ jobs:
            type=raw,value=latest,enable=${{ github.ref == format('refs/heads/{0}', 'main') }}

      - name: Login to Docker Registry
-        uses: docker/login-action@v2
+        uses: docker/login-action@v3
        with:
          registry: ${{ env.PILOT_IMAGE_REGISTRY }}
          username: ${{ secrets.REGISTRY_USERNAME }}
@@ -169,7 +169,7 @@ jobs:
      GATEWAY_IMAGE_NAME: ${{ vars.GATEWAY_IMAGE_NAME || 'higress/gateway' }}
    steps:
      - name: "Checkout ${{ github.ref }}"
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
        with:
          fetch-depth: 1

@@ -184,12 +184,12 @@ jobs:
          swap-storage: true

      - name: "Setup Go"
-        uses: actions/setup-go@v3
+        uses: actions/setup-go@v5
        with:
          go-version: 1.19

      - name: Setup Golang Caches
-        uses: actions/cache@v3
+        uses: actions/cache@v4
        with:
          path: |-
            ~/.cache/go-build
@@ -198,7 +198,7 @@ jobs:
          restore-keys: ${{ runner.os }}-go

      - name: Setup Submodule Caches
-        uses: actions/cache@v3
+        uses: actions/cache@v4
        with:
          path: |-
            envoy
@@ -209,7 +209,7 @@ jobs:

      - name: Calculate Docker metadata
        id: docker-meta
-        uses: docker/metadata-action@v4
+        uses: docker/metadata-action@v5
        with:
          images: |
            ${{ env.GATEWAY_IMAGE_REGISTRY }}/${{ env.GATEWAY_IMAGE_NAME }}
@@ -220,7 +220,7 @@ jobs:
            type=raw,value=latest,enable=${{ github.ref == format('refs/heads/{0}', 'main') }}

      - name: Login to Docker Registry
-        uses: docker/login-action@v2
+        uses: docker/login-action@v3
        with:
          registry: ${{ env.GATEWAY_IMAGE_REGISTRY }}
          username: ${{ secrets.REGISTRY_USERNAME }}
--- a/.github/workflows/codeql-analysis.yaml
+++ b/.github/workflows/codeql-analysis.yaml
@@ -34,11 +34,11 @@ jobs:
    steps:
      # step 1
      - name: "Checkout repository"
-        uses: actions/checkout@v2
+        uses: actions/checkout@v4

      # step 2: Initializes the CodeQL tools for scanning.
      - name: "Initialize CodeQL"
-        uses: github/codeql-action/init@v1
+        uses: github/codeql-action/init@v2
        with:
          languages: ${{ matrix.language }}
          # If you wish to specify custom queries, you can do so here or in a config file.
@@ -50,7 +50,7 @@ jobs:
      # Autobuild attempts to build any compiled languages  (C/C++, C#, or Java).
      # If this step fails, then you should remove it and run the build manually (see below)
      - name: "Autobuild"
-        uses: github/codeql-action/autobuild@v1
+        uses: github/codeql-action/autobuild@v2

      # step 4
      # ℹ️ Command-line programs to run using the OS shell.
@@ -66,4 +66,4 @@ jobs:

      # step 5
      - name: "Perform CodeQL Analysis"
-        uses: github/codeql-action/analyze@v1
+        uses: github/codeql-action/analyze@v2
--- a/.github/workflows/deploy-standalone-to-oss.yaml
+++ b/.github/workflows/deploy-standalone-to-oss.yaml
@@ -14,7 +14,7 @@ jobs:
    steps:
      # Step 1
      - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
      # Step 2
      - id: package
        name: Prepare Standalone Package
--- a/.github/workflows/deploy-to-oss.yaml
+++ b/.github/workflows/deploy-to-oss.yaml
@@ -14,7 +14,7 @@ jobs:
    steps:
      # Step 1
      - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
      # Step 2
      - name: Download Helm Charts Index
        uses: doggycool/ossutil-github-action@master
--- a/.github/workflows/latest-release.yaml
+++ b/.github/workflows/latest-release.yaml
@@ -9,7 +9,7 @@ jobs:
  latest-release:
    runs-on: ubuntu-latest
    steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4

    - name: Build hgctl latest multiarch binaries
      run: | 
@@ -46,7 +46,7 @@ jobs:
        GITHUB_REPOSITORY: ${{ github.repository_owner }}/${{ github.event.repository.name }}

    - name: Recreate the Latest Release and Tag
-      uses: softprops/action-gh-release@v1
+      uses: softprops/action-gh-release@v2
      with:
        draft: false
        prerelease: true
--- a/.github/workflows/license-checker.yaml
+++ b/.github/workflows/license-checker.yaml
@@ -10,7 +10,7 @@ jobs:
    steps:
      # step 1
      - name: Checkout
-        uses: actions/checkout@v2.4.0
+        uses: actions/checkout@v4
      # step 2
      - name: Check License Header
        uses: apache/skywalking-eyes/header@25edfc2fd8d52fb266653fb5f6c42da633d85c07
@@ -24,4 +24,4 @@ jobs:
        with:
          log: info
          config: .licenserc.yaml
-          mode: check
+          mode: check
--- a/.github/workflows/release-hgctl.yaml
+++ b/.github/workflows/release-hgctl.yaml
@@ -12,7 +12,7 @@ jobs:
    env:
      HGCTL_VERSION: ${{github.ref_name}}
    steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4

    - name: Build hgctl latest multiarch binaries
      run: |
@@ -25,7 +25,7 @@ jobs:
        zip -q -r hgctl_${{ env.HGCTL_VERSION }}_windows_arm64.zip out/windows_arm64/

    - name: Upload hgctl packages to the GitHub release
-      uses: softprops/action-gh-release@v1
+      uses: softprops/action-gh-release@v2
      if: startsWith(github.ref, 'refs/tags/')
      with:
        files: |
@@ -34,4 +34,4 @@ jobs:
          hgctl_${{ env.HGCTL_VERSION }}_darwin_amd64.tar.gz
          hgctl_${{ env.HGCTL_VERSION }}_darwin_arm64.tar.gz
          hgctl_${{ env.HGCTL_VERSION }}_windows_amd64.zip
-          hgctl_${{ env.HGCTL_VERSION }}_windows_arm64.zip
+          hgctl_${{ env.HGCTL_VERSION }}_windows_arm64.zip
--- a/Makefile.core.mk
+++ b/Makefile.core.mk
@@ -138,11 +138,11 @@ export ENVOY_TAR_PATH:=/home/package/envoy.tar.gz

 external/package/envoy-amd64.tar.gz:
 #	cd external/proxy; BUILD_WITH_CONTAINER=1  make test_release
-	cd external/package; wget -O envoy-amd64.tar.gz "https://github.com/alibaba/higress/releases/download/v1.4.0-rc.1/envoy-symbol-amd64.tar.gz"
+	cd external/package; wget -O envoy-amd64.tar.gz "https://github.com/alibaba/higress/releases/download/v1.4.1/envoy-symbol-amd64.tar.gz"

 external/package/envoy-arm64.tar.gz:
 #	cd external/proxy; BUILD_WITH_CONTAINER=1  make test_release
-	cd external/package; wget -O envoy-arm64.tar.gz "https://github.com/alibaba/higress/releases/download/v1.4.0-rc.1/envoy-symbol-arm64.tar.gz"
+	cd external/package; wget -O envoy-arm64.tar.gz "https://github.com/alibaba/higress/releases/download/v1.4.1/envoy-symbol-arm64.tar.gz"

 build-pilot:
 	cd external/istio; rm -rf out/linux_amd64; GOOS_LOCAL=linux TARGET_OS=linux TARGET_ARCH=amd64 BUILD_WITH_CONTAINER=1 make build-linux
@@ -177,8 +177,8 @@ install: pre-install
 	cd helm/higress; helm dependency build
 	helm install higress helm/higress -n higress-system --create-namespace --set 'global.local=true'

-ENVOY_LATEST_IMAGE_TAG ?= sha-93966bf
-ISTIO_LATEST_IMAGE_TAG ?= sha-b00f79f
+ENVOY_LATEST_IMAGE_TAG ?= sha-59acb61
+ISTIO_LATEST_IMAGE_TAG ?= sha-59acb61

 install-dev: pre-install
 	helm install higress helm/core -n higress-system --create-namespace --set 'controller.tag=$(TAG)' --set 'gateway.replicas=1' --set 'pilot.tag=$(ISTIO_LATEST_IMAGE_TAG)' --set 'gateway.tag=$(ENVOY_LATEST_IMAGE_TAG)' --set 'global.local=true'
--- a/README.md
+++ b/README.md
@@ -1,17 +1,17 @@
 <h1 align="center">
    <img src="https://img.alicdn.com/imgextra/i2/O1CN01NwxLDd20nxfGBjxmZ_!!6000000006895-2-tps-960-290.png" alt="Higress" width="240" height="72.5">
  <br>
-  Cloud Native API Gateway
+  AI Native API Gateway
 </h1>

 [![Build Status](https://github.com/alibaba/higress/actions/workflows/build-and-test.yaml/badge.svg?branch=main)](https://github.com/alibaba/higress/actions)
 [![license](https://img.shields.io/github/license/alibaba/higress.svg)](https://www.apache.org/licenses/LICENSE-2.0.html)

 [**官网**](https://higress.io/) &nbsp; |
-&nbsp; [**文档**](https://higress.io/zh-cn/docs/overview/what-is-higress) &nbsp; |
-&nbsp; [**博客**](https://higress.io/zh-cn/blog) &nbsp; |
-&nbsp; [**开发指引**](https://higress.io/zh-cn/docs/developers/developers_dev) &nbsp; |
-&nbsp; [**Higress 企业版**](https://www.aliyun.com/product/aliware/mse?spm=higress-website.topbar.0.0.0) &nbsp;
+&nbsp; [**文档**](https://higress.io/docs/latest/user/quickstart/) &nbsp; |
+&nbsp; [**博客**](https://higress.io/blog/) &nbsp; |
+&nbsp; [**开发指引**](https://higress.io/docs/latest/dev/architecture/) &nbsp; |
+&nbsp; [**AI插件**](https://higress.io/plugin/) &nbsp;


 <p>
@@ -19,9 +19,15 @@
 </p>


-Higress 是基于阿里内部两年多的 Envoy Gateway 实践沉淀，以开源 [Istio](https://github.com/istio/istio) 与 [Envoy](https://github.com/envoyproxy/envoy) 为核心构建的云原生 API 网关。Higress 实现了安全防护网关、流量网关、微服务网关三层网关合一，可以显著降低网关的部署和运维成本。
+Higress 是基于阿里内部多年的 Envoy Gateway 实践沉淀，以开源 [Istio](https://github.com/istio/istio) 与 [Envoy](https://github.com/envoyproxy/envoy) 为核心构建的云原生 API 网关。
+
+Higress 是面向 AI 原生设计的 API 网关，在阿里内部，承载了通义千问 APP、百炼大模型 API、机器学习 PAI 平台等 AI 业务的流量。
+
+Higress 能够用统一的协议对接国内外所有 LLM 模型厂商，同时具备丰富的 AI 可观测、多模型负载均衡/fallback、AI token 流控、AI 缓存等能力：
+
+![](https://img.alicdn.com/imgextra/i1/O1CN01fNnhCp1cV8mYPRFeS_!!6000000003605-0-tps-1080-608.jpg)
+

-![arch](https://img.alicdn.com/imgextra/i1/O1CN01iO9ph825juHbOIg75_!!6000000007563-2-tps-2483-2024.png)

 ## Summary
    
@@ -34,6 +40,10 @@ Higress 是基于阿里内部两年多的 Envoy Gateway 实践沉淀，以开源

 ## 使用场景

+- **AI 网关**:
+
+  Higress 提供了一站式的 AI 插件集，可以增强依赖 AI 能力业务的稳定性、灵活性、可观测性，使得业务与 AI 的集成更加便捷和高效。
+
 - **Kubernetes Ingress 网关**:

  Higress 可以作为 K8s 集群的 Ingress 入口网关, 并且兼容了大量 K8s Nginx Ingress 的注解，可以从 K8s Nginx Ingress 快速平滑迁移到 Higress。
@@ -56,27 +66,36 @@ Higress 是基于阿里内部两年多的 Envoy Gateway 实践沉淀，以开源

  脱胎于阿里巴巴2年多生产验证的内部产品，支持每秒请求量达数十万级的大规模场景。

-  彻底摆脱 reload 引起的流量抖动，配置变更毫秒级生效且业务无感。
-  
- **平滑演进**
+  彻底摆脱 Nginx reload 引起的流量抖动，配置变更毫秒级生效且业务无感。对 AI 业务等长连接场景特别友好。

-  支持 Nacos/Zookeeper/Eureka 等多种注册中心，可以不依赖 K8s Service 进行服务发现，支持非容器架构平滑演进到云原生架构。
+- **流式处理**

-  支持从 Nginx Ingress Controller 平滑迁移，支持平滑过渡到 Gateway API，支持业务架构平滑演进到 ServiceMesh。
+  支持真正的完全流式处理请求/响应 Body，Wasm 插件很方便地自定义处理 SSE （Server-Sent Events）等流式协议的报文。

- **兼收并蓄**
-  
-  兼容 Nginx Ingress Annotation 80%+ 的使用场景，且提供功能更丰富的 Higress Annotation 注解。
-  
-  兼容 Ingress API/Gateway API/Istio API，可以组合多种 CRD 实现流量精细化管理。
-  
+  在 AI 业务等大带宽场景下，可以显著降低内存开销。  
+    
 - **便于扩展**
  
-  提供 Wasm、Lua、进程外三种插件扩展机制，支持多语言编写插件，生效粒度支持全局级、域名级，路由级。
+  提供丰富的官方插件库，涵盖 AI、流量管理、安全防护等常用功能，满足90%以上的业务场景需求。
+
+  主打 Wasm 插件扩展，通过沙箱隔离确保内存安全，支持多种编程语言，允许插件版本独立升级，实现流量无损热更新网关逻辑。
+
+- **安全易用**
+  
+  基于 Ingress API 和 Gateway API 标准，提供开箱即用的 UI 控制台，WAF 防护插件、IP/Cookie CC 防护插件开箱即用。
+
+  支持对接 Let's Encrypt 自动签发和续签免费证书，并且可以脱离 K8s 部署，一行 Docker 命令即可启动，方便个人开发者使用。

-  插件支持热更新，变更插件逻辑和配置都对流量无损。

 ## 功能展示
+
+### AI 网关 Demo 展示
+
+[从 OpenAI 到其他大模型，30 秒完成迁移
+](https://www.bilibili.com/video/BV1dT421a7w7/?spm_id_from=333.788.recommend_more_video.14)
+
+
+### Higress UI 控制台
    
 - **丰富的可观测**

--- a/2
+++ b/2
@@ -1 +1 @@
-v1.4.0
+v1.4.2
--- a/api/extensions/v1alpha1/wasm.pb.go
+++ b/api/extensions/v1alpha1/wasm.pb.go
@@ -301,6 +301,7 @@ type MatchRule struct {
 	Domain               []string      `protobuf:"bytes,2,rep,name=domain,proto3" json:"domain,omitempty"`
 	Config               *types.Struct `protobuf:"bytes,3,opt,name=config,proto3" json:"config,omitempty"`
 	ConfigDisable        bool          `protobuf:"varint,4,opt,name=config_disable,json=configDisable,proto3" json:"config_disable,omitempty"`
+	Service              []string      `protobuf:"bytes,5,rep,name=service,proto3" json:"service,omitempty"`
 	XXX_NoUnkeyedLiteral struct{}      `json:"-"`
 	XXX_unrecognized     []byte        `json:"-"`
 	XXX_sizecache        int32         `json:"-"`
@@ -367,6 +368,13 @@ func (m *MatchRule) GetConfigDisable() bool {
 	return false
 }

+func (m *MatchRule) GetService() []string {
+	if m != nil {
+		return m.Service
+	}
+	return nil
+}
+
 func init() {
 	proto.RegisterEnum("higress.extensions.v1alpha1.PluginPhase", PluginPhase_name, PluginPhase_value)
 	proto.RegisterEnum("higress.extensions.v1alpha1.PullPolicy", PullPolicy_name, PullPolicy_value)
@@ -377,46 +385,47 @@ func init() {
 func init() { proto.RegisterFile("extensions/v1alpha1/wasm.proto", fileDescriptor_4d60b240916c4e18) }

 var fileDescriptor_4d60b240916c4e18 = []byte{
-	// 619 bytes of a gzipped FileDescriptorProto
-	0x1f, 0x8b, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0xff, 0x7c, 0x94, 0xdd, 0x4e, 0xdb, 0x4c,
-	0x10, 0x86, 0x71, 0x02, 0x81, 0x4c, 0x80, 0xcf, 0xac, 0xbe, 0xd2, 0x15, 0x54, 0x69, 0x84, 0xd4,
-	0xd6, 0xe5, 0xc0, 0x16, 0xa1, 0x3f, 0x27, 0x15, 0x6a, 0x80, 0xb4, 0x44, 0x6d, 0x53, 0xcb, 0x86,
-	0x56, 0xe5, 0xc4, 0xda, 0x98, 0x8d, 0xb3, 0xea, 0xfa, 0x47, 0xde, 0x35, 0x34, 0x17, 0xd2, 0x7b,
-	0xea, 0x61, 0x2f, 0xa1, 0xe2, 0x2e, 0x7a, 0x56, 0x65, 0x6d, 0x43, 0x42, 0xab, 0x9c, 0xed, 0xce,
-	0x3c, 0x33, 0xf3, 0xbe, 0xe3, 0x95, 0xa1, 0x49, 0xbf, 0x49, 0x1a, 0x09, 0x16, 0x47, 0xc2, 0xba,
-	0xdc, 0x23, 0x3c, 0x19, 0x91, 0x3d, 0xeb, 0x8a, 0x88, 0xd0, 0x4c, 0xd2, 0x58, 0xc6, 0x68, 0x7b,
-	0xc4, 0x82, 0x94, 0x0a, 0x61, 0xde, 0x72, 0x66, 0xc9, 0x6d, 0x35, 0x83, 0x38, 0x0e, 0x38, 0xb5,
-	0x14, 0x3a, 0xc8, 0x86, 0xd6, 0x55, 0x4a, 0x92, 0x84, 0xa6, 0x22, 0x2f, 0xde, 0x7a, 0x70, 0x37,
-	0x2f, 0x64, 0x9a, 0xf9, 0x32, 0xcf, 0xee, 0xfc, 0x5e, 0x04, 0xf8, 0x4c, 0x44, 0x68, 0xf3, 0x2c,
-	0x60, 0x11, 0xd2, 0xa1, 0x9a, 0xa5, 0x1c, 0x57, 0x5a, 0x9a, 0x51, 0x77, 0x26, 0x47, 0xb4, 0x09,
-	0x35, 0x31, 0x22, 0xed, 0xe7, 0x2f, 0x70, 0x55, 0x05, 0x8b, 0x1b, 0x72, 0x61, 0x83, 0x85, 0x24,
-	0xa0, 0x5e, 0x92, 0x71, 0xee, 0x25, 0x31, 0x67, 0xfe, 0x18, 0x2f, 0xb6, 0x34, 0x63, 0xbd, 0xfd,
-	0xc4, 0x9c, 0xa3, 0xd7, 0xb4, 0x33, 0xce, 0x6d, 0x85, 0x3b, 0xff, 0xa9, 0x0e, 0xb7, 0x01, 0xb4,
-	0x3b, 0xd3, 0x54, 0x50, 0x3f, 0xa5, 0x12, 0x2f, 0xa9, 0xb9, 0xb7, 0xac, 0xab, 0xc2, 0xe8, 0x29,
-	0xe8, 0x97, 0x34, 0x65, 0x43, 0xe6, 0x13, 0xc9, 0xe2, 0xc8, 0xfb, 0x4a, 0xc7, 0xb8, 0x96, 0xa3,
-	0xd3, 0xf1, 0x77, 0x74, 0x8c, 0x5e, 0xc1, 0x5a, 0xa2, 0xfc, 0x79, 0x7e, 0x1c, 0x0d, 0x59, 0x80,
-	0x97, 0x5b, 0x9a, 0xd1, 0x68, 0xdf, 0x37, 0xf3, 0xd5, 0x98, 0xe5, 0x6a, 0x4c, 0x57, 0xad, 0xc6,
-	0x59, 0xcd, 0xe9, 0x23, 0x05, 0xa3, 0x87, 0xd0, 0x28, 0xaa, 0x23, 0x12, 0x52, 0xbc, 0xa2, 0x66,
-	0x40, 0x1e, 0xea, 0x93, 0x90, 0xa2, 0x03, 0x58, 0x4a, 0x46, 0x44, 0x50, 0x5c, 0x57, 0xf6, 0x8d,
-	0xf9, 0xf6, 0x55, 0x9d, 0x3d, 0xe1, 0x9d, 0xbc, 0x0c, 0xbd, 0x84, 0x95, 0x24, 0x65, 0x71, 0xca,
-	0xe4, 0x18, 0x83, 0x52, 0xb6, 0xfd, 0x97, 0xb2, 0x5e, 0x24, 0xf7, 0xdb, 0x9f, 0x08, 0xcf, 0xa8,
-	0x73, 0x03, 0xa3, 0x03, 0x58, 0xbf, 0xa0, 0x43, 0x92, 0x71, 0x59, 0x1a, 0xa3, 0xf3, 0x8d, 0xad,
-	0x15, 0x78, 0xe1, 0xec, 0x2d, 0x34, 0x42, 0x22, 0xfd, 0x91, 0x97, 0x66, 0x9c, 0x0a, 0x3c, 0x6c,
-	0x55, 0x8d, 0x46, 0xfb, 0xf1, 0x5c, 0xf9, 0x1f, 0x26, 0xbc, 0x93, 0x71, 0xea, 0x40, 0x58, 0x1e,
-	0x05, 0x7a, 0x06, 0x9b, 0xb3, 0x42, 0xbc, 0x0b, 0x26, 0xc8, 0x80, 0x53, 0x1c, 0xb4, 0x34, 0x63,
-	0xc5, 0xf9, 0x7f, 0x66, 0xee, 0x71, 0x9e, 0xdb, 0xf9, 0xae, 0x41, 0xfd, 0xa6, 0x1f, 0xc2, 0xb0,
-	0xcc, 0x22, 0x35, 0x18, 0x6b, 0xad, 0xaa, 0x51, 0x77, 0xca, 0xeb, 0xe4, 0x09, 0x5e, 0xc4, 0x21,
-	0x61, 0x11, 0xae, 0xa8, 0x44, 0x71, 0x43, 0x16, 0xd4, 0x0a, 0xdb, 0xd5, 0xf9, 0xb6, 0x0b, 0x0c,
-	0x3d, 0x82, 0xf5, 0x3b, 0xf2, 0x16, 0x95, 0xbc, 0x35, 0x7f, 0x5a, 0xd7, 0x6e, 0x17, 0x1a, 0x53,
-	0x5f, 0x09, 0xdd, 0x83, 0x8d, 0xb3, 0xbe, 0x6b, 0x77, 0x8f, 0x7a, 0x6f, 0x7a, 0xdd, 0x63, 0xcf,
-	0x3e, 0xe9, 0xb8, 0x5d, 0x7d, 0x01, 0xd5, 0x61, 0xa9, 0x73, 0x76, 0x7a, 0xd2, 0xd7, 0xb5, 0xf2,
-	0x78, 0xae, 0x57, 0x26, 0x47, 0xf7, 0xb4, 0x73, 0xea, 0xea, 0xd5, 0xdd, 0x43, 0x80, 0xa9, 0xa7,
-	0xbd, 0x09, 0x68, 0xa6, 0xcb, 0xc7, 0xf7, 0xbd, 0xa3, 0x2f, 0xfa, 0x02, 0xd2, 0x61, 0xb5, 0x37,
-	0xec, 0xc7, 0xd2, 0x4e, 0xa9, 0xa0, 0x91, 0xd4, 0x35, 0x04, 0x50, 0xeb, 0xf0, 0x2b, 0x32, 0x16,
-	0x7a, 0xe5, 0xf0, 0xf5, 0x8f, 0xeb, 0xa6, 0xf6, 0xf3, 0xba, 0xa9, 0xfd, 0xba, 0x6e, 0x6a, 0xe7,
-	0xed, 0x80, 0xc9, 0x51, 0x36, 0x30, 0xfd, 0x38, 0xb4, 0x08, 0x67, 0x03, 0x32, 0x20, 0x56, 0xf1,
-	0xb1, 0x2c, 0x92, 0x30, 0xeb, 0x1f, 0xbf, 0x91, 0x41, 0x4d, 0x2d, 0x63, 0xff, 0x4f, 0x00, 0x00,
-	0x00, 0xff, 0xff, 0xb9, 0xf2, 0x67, 0xbe, 0x64, 0x04, 0x00, 0x00,
+	// 631 bytes of a gzipped FileDescriptorProto
+	0x1f, 0x8b, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0xff, 0x7c, 0x94, 0xdd, 0x6e, 0xd3, 0x4c,
+	0x10, 0x86, 0xeb, 0xa4, 0x49, 0x9b, 0x49, 0xdb, 0xcf, 0x5d, 0x7d, 0x94, 0x55, 0x8b, 0x42, 0x54,
+	0x09, 0x30, 0x3d, 0xb0, 0xd5, 0x94, 0x9f, 0x13, 0x54, 0x91, 0xb6, 0x81, 0x46, 0x40, 0xb0, 0xec,
+	0x16, 0x44, 0x4f, 0xac, 0x8d, 0xbb, 0x71, 0x56, 0xac, 0x7f, 0xe4, 0x5d, 0xb7, 0xe4, 0xaa, 0xb8,
+	0x0d, 0x0e, 0xb9, 0x04, 0xd4, 0xbb, 0xe0, 0x0c, 0x65, 0xed, 0x34, 0x49, 0x41, 0x39, 0xdb, 0x9d,
+	0x79, 0x66, 0xe6, 0x7d, 0xc7, 0x2b, 0x43, 0x83, 0x7e, 0x93, 0x34, 0x12, 0x2c, 0x8e, 0x84, 0x75,
+	0xb5, 0x4f, 0x78, 0x32, 0x24, 0xfb, 0xd6, 0x35, 0x11, 0xa1, 0x99, 0xa4, 0xb1, 0x8c, 0xd1, 0xce,
+	0x90, 0x05, 0x29, 0x15, 0xc2, 0x9c, 0x72, 0xe6, 0x84, 0xdb, 0x6e, 0x04, 0x71, 0x1c, 0x70, 0x6a,
+	0x29, 0xb4, 0x9f, 0x0d, 0xac, 0xeb, 0x94, 0x24, 0x09, 0x4d, 0x45, 0x5e, 0xbc, 0xfd, 0xe0, 0x6e,
+	0x5e, 0xc8, 0x34, 0xf3, 0x65, 0x9e, 0xdd, 0xfd, 0xbd, 0x0c, 0xf0, 0x99, 0x88, 0xd0, 0xe6, 0x59,
+	0xc0, 0x22, 0xa4, 0x43, 0x39, 0x4b, 0x39, 0x2e, 0x35, 0x35, 0xa3, 0xe6, 0x8c, 0x8f, 0x68, 0x0b,
+	0xaa, 0x62, 0x48, 0x5a, 0xcf, 0x5f, 0xe0, 0xb2, 0x0a, 0x16, 0x37, 0xe4, 0xc2, 0x26, 0x0b, 0x49,
+	0x40, 0xbd, 0x24, 0xe3, 0xdc, 0x4b, 0x62, 0xce, 0xfc, 0x11, 0x5e, 0x6e, 0x6a, 0xc6, 0x46, 0xeb,
+	0x89, 0xb9, 0x40, 0xaf, 0x69, 0x67, 0x9c, 0xdb, 0x0a, 0x77, 0xfe, 0x53, 0x1d, 0xa6, 0x01, 0xb4,
+	0x37, 0xd7, 0x54, 0x50, 0x3f, 0xa5, 0x12, 0x57, 0xd4, 0xdc, 0x29, 0xeb, 0xaa, 0x30, 0x7a, 0x0a,
+	0xfa, 0x15, 0x4d, 0xd9, 0x80, 0xf9, 0x44, 0xb2, 0x38, 0xf2, 0xbe, 0xd2, 0x11, 0xae, 0xe6, 0xe8,
+	0x6c, 0xfc, 0x1d, 0x1d, 0xa1, 0x57, 0xb0, 0x9e, 0x28, 0x7f, 0x9e, 0x1f, 0x47, 0x03, 0x16, 0xe0,
+	0x95, 0xa6, 0x66, 0xd4, 0x5b, 0xf7, 0xcd, 0x7c, 0x35, 0xe6, 0x64, 0x35, 0xa6, 0xab, 0x56, 0xe3,
+	0xac, 0xe5, 0xf4, 0xb1, 0x82, 0xd1, 0x43, 0xa8, 0x17, 0xd5, 0x11, 0x09, 0x29, 0x5e, 0x55, 0x33,
+	0x20, 0x0f, 0xf5, 0x48, 0x48, 0xd1, 0x21, 0x54, 0x92, 0x21, 0x11, 0x14, 0xd7, 0x94, 0x7d, 0x63,
+	0xb1, 0x7d, 0x55, 0x67, 0x8f, 0x79, 0x27, 0x2f, 0x43, 0x2f, 0x61, 0x35, 0x49, 0x59, 0x9c, 0x32,
+	0x39, 0xc2, 0xa0, 0x94, 0xed, 0xfc, 0xa5, 0xac, 0x1b, 0xc9, 0x83, 0xd6, 0x27, 0xc2, 0x33, 0xea,
+	0xdc, 0xc2, 0xe8, 0x10, 0x36, 0x2e, 0xe9, 0x80, 0x64, 0x5c, 0x4e, 0x8c, 0xd1, 0xc5, 0xc6, 0xd6,
+	0x0b, 0xbc, 0x70, 0xf6, 0x16, 0xea, 0x21, 0x91, 0xfe, 0xd0, 0x4b, 0x33, 0x4e, 0x05, 0x1e, 0x34,
+	0xcb, 0x46, 0xbd, 0xf5, 0x78, 0xa1, 0xfc, 0x0f, 0x63, 0xde, 0xc9, 0x38, 0x75, 0x20, 0x9c, 0x1c,
+	0x05, 0x7a, 0x06, 0x5b, 0xf3, 0x42, 0xbc, 0x4b, 0x26, 0x48, 0x9f, 0x53, 0x1c, 0x34, 0x35, 0x63,
+	0xd5, 0xf9, 0x7f, 0x6e, 0xee, 0x49, 0x9e, 0xdb, 0xfd, 0xae, 0x41, 0xed, 0xb6, 0x1f, 0xc2, 0xb0,
+	0xc2, 0x22, 0x35, 0x18, 0x6b, 0xcd, 0xb2, 0x51, 0x73, 0x26, 0xd7, 0xf1, 0x13, 0xbc, 0x8c, 0x43,
+	0xc2, 0x22, 0x5c, 0x52, 0x89, 0xe2, 0x86, 0x2c, 0xa8, 0x16, 0xb6, 0xcb, 0x8b, 0x6d, 0x17, 0x18,
+	0x7a, 0x04, 0x1b, 0x77, 0xe4, 0x2d, 0x2b, 0x79, 0xeb, 0xfe, 0xac, 0xae, 0xb1, 0x12, 0x41, 0xd3,
+	0x2b, 0xe6, 0x53, 0x5c, 0xc9, 0x95, 0x14, 0xd7, 0xbd, 0x0e, 0xd4, 0x67, 0xbe, 0x1f, 0xba, 0x07,
+	0x9b, 0xe7, 0x3d, 0xd7, 0xee, 0x1c, 0x77, 0xdf, 0x74, 0x3b, 0x27, 0x9e, 0x7d, 0xda, 0x76, 0x3b,
+	0xfa, 0x12, 0xaa, 0x41, 0xa5, 0x7d, 0x7e, 0x76, 0xda, 0xd3, 0xb5, 0xc9, 0xf1, 0x42, 0x2f, 0x8d,
+	0x8f, 0xee, 0x59, 0xfb, 0xcc, 0xd5, 0xcb, 0x7b, 0x47, 0x00, 0x33, 0x8f, 0x7e, 0x0b, 0xd0, 0x5c,
+	0x97, 0x8f, 0xef, 0xbb, 0xc7, 0x5f, 0xf4, 0x25, 0xa4, 0xc3, 0x5a, 0x77, 0xd0, 0x8b, 0xa5, 0x9d,
+	0x52, 0x41, 0x23, 0xa9, 0x6b, 0x08, 0xa0, 0xda, 0xe6, 0xd7, 0x64, 0x24, 0xf4, 0xd2, 0xd1, 0xeb,
+	0x1f, 0x37, 0x0d, 0xed, 0xe7, 0x4d, 0x43, 0xfb, 0x75, 0xd3, 0xd0, 0x2e, 0x5a, 0x01, 0x93, 0xc3,
+	0xac, 0x6f, 0xfa, 0x71, 0x68, 0x11, 0xce, 0xfa, 0xa4, 0x4f, 0xac, 0xe2, 0x33, 0x5a, 0x24, 0x61,
+	0xd6, 0x3f, 0x7e, 0x30, 0xfd, 0xaa, 0x5a, 0xd3, 0xc1, 0x9f, 0x00, 0x00, 0x00, 0xff, 0xff, 0x0b,
+	0x3c, 0xc3, 0xcf, 0x7e, 0x04, 0x00, 0x00,
 }

 func (m *WasmPlugin) Marshal() (dAtA []byte, err error) {
@@ -581,6 +590,15 @@ func (m *MatchRule) MarshalToSizedBuffer(dAtA []byte) (int, error) {
 		i -= len(m.XXX_unrecognized)
 		copy(dAtA[i:], m.XXX_unrecognized)
 	}
+	if len(m.Service) > 0 {
+		for iNdEx := len(m.Service) - 1; iNdEx >= 0; iNdEx-- {
+			i -= len(m.Service[iNdEx])
+			copy(dAtA[i:], m.Service[iNdEx])
+			i = encodeVarintWasm(dAtA, i, uint64(len(m.Service[iNdEx])))
+			i--
+			dAtA[i] = 0x2a
+		}
+	}
 	if m.ConfigDisable {
 		i--
 		if m.ConfigDisable {
@@ -719,6 +737,12 @@ func (m *MatchRule) Size() (n int) {
 	if m.ConfigDisable {
 		n += 2
 	}
+	if len(m.Service) > 0 {
+		for _, s := range m.Service {
+			l = len(s)
+			n += 1 + l + sovWasm(uint64(l))
+		}
+	}
 	if m.XXX_unrecognized != nil {
 		n += len(m.XXX_unrecognized)
 	}
@@ -1291,6 +1315,38 @@ func (m *MatchRule) Unmarshal(dAtA []byte) error {
 				}
 			}
 			m.ConfigDisable = bool(v != 0)
+		case 5:
+			if wireType != 2 {
+				return fmt.Errorf("proto: wrong wireType = %d for field Service", wireType)
+			}
+			var stringLen uint64
+			for shift := uint(0); ; shift += 7 {
+				if shift >= 64 {
+					return ErrIntOverflowWasm
+				}
+				if iNdEx >= l {
+					return io.ErrUnexpectedEOF
+				}
+				b := dAtA[iNdEx]
+				iNdEx++
+				stringLen |= uint64(b&0x7F) << shift
+				if b < 0x80 {
+					break
+				}
+			}
+			intStringLen := int(stringLen)
+			if intStringLen < 0 {
+				return ErrInvalidLengthWasm
+			}
+			postIndex := iNdEx + intStringLen
+			if postIndex < 0 {
+				return ErrInvalidLengthWasm
+			}
+			if postIndex > l {
+				return io.ErrUnexpectedEOF
+			}
+			m.Service = append(m.Service, string(dAtA[iNdEx:postIndex]))
+			iNdEx = postIndex
 		default:
 			iNdEx = preIndex
 			skippy, err := skipWasm(dAtA[iNdEx:])
--- a/api/extensions/v1alpha1/wasm.proto
+++ b/api/extensions/v1alpha1/wasm.proto
@@ -114,6 +114,7 @@ message MatchRule {
  repeated string domain = 2;
  google.protobuf.Struct config = 3;
  bool config_disable = 4;
+  repeated string service = 5;
 }

 // The phase in the filter chain where the plugin will be injected.
--- a/api/kubernetes/customresourcedefinitions.gen.yaml
+++ b/api/kubernetes/customresourcedefinitions.gen.yaml
@@ -64,6 +64,10 @@ spec:
                      items:
                        type: string
                      type: array
+                    service:
+                      items:
+                        type: string
+                      type: array
                  type: object
                type: array
              phase:
--- a/cmd/hgctl/config/gateway_config.go
+++ b/cmd/hgctl/config/gateway_config.go
@@ -22,6 +22,7 @@ import (

 	"github.com/alibaba/higress/pkg/cmd/hgctl/kubernetes"
 	"github.com/alibaba/higress/pkg/cmd/options"
+	"istio.io/istio/istioctl/pkg/writer/envoy/configdump"
 	"k8s.io/apimachinery/pkg/types"
 	"sigs.k8s.io/yaml"
 )
@@ -61,6 +62,23 @@ func NewDefaultGetEnvoyConfigOptions() *GetEnvoyConfigOptions {
 	}
 }

+func setupConfigdumpEnvoyConfigWriter(debug []byte, stdout io.Writer) (*configdump.ConfigWriter, error) {
+	cw := &configdump.ConfigWriter{Stdout: stdout}
+	err := cw.Prime(debug)
+	if err != nil {
+		return nil, err
+	}
+	return cw, nil
+}
+
+func GetEnvoyConfigWriter(config *GetEnvoyConfigOptions, stdout io.Writer) (*configdump.ConfigWriter, error) {
+	configDump, err := retrieveConfigDump(config.PodName, config.PodNamespace, config.BindAddress, config.IncludeEds)
+	if err != nil {
+		return nil, err
+	}
+	return setupConfigdumpEnvoyConfigWriter(configDump, stdout)
+}
+
 func GetEnvoyConfig(config *GetEnvoyConfigOptions) ([]byte, error) {
 	configDump, err := retrieveConfigDump(config.PodName, config.PodNamespace, config.BindAddress, config.IncludeEds)
 	if err != nil {
@@ -144,14 +162,12 @@ func formatGatewayConfig(configDump any, output string) ([]byte, error) {
 	if err != nil {
 		return nil, err
 	}
-
 	if output == "yaml" {
 		out, err = yaml.JSONToYAML(out)
 		if err != nil {
 			return nil, err
 		}
 	}
-
 	return out, nil
 }

--- a/envoy/1.20/patches/envoy/20240610-optimize-xds.patch
+++ b/envoy/1.20/patches/envoy/20240610-optimize-xds.patch
@@ -0,0 +1,259 @@
+diff --git a/source/common/router/BUILD b/source/common/router/BUILD
+index 5c58501..4db76cd 100644
+--- a/source/common/router/BUILD
+++ b/source/common/router/BUILD
+@@ -212,6 +212,7 @@ envoy_cc_library(
+         "//envoy/router:rds_interface",
+         "//envoy/router:scopes_interface",
+         "//envoy/thread_local:thread_local_interface",
+        "//source/common/protobuf:utility_lib",
+         "@envoy_api//envoy/config/route/v3:pkg_cc_proto",
+         "@envoy_api//envoy/extensions/filters/network/http_connection_manager/v3:pkg_cc_proto",
+     ],
+diff --git a/source/common/router/config_impl.cc b/source/common/router/config_impl.cc
+index ff7b4c8..5ac4523 100644
+--- a/source/common/router/config_impl.cc
+++ b/source/common/router/config_impl.cc
+@@ -550,19 +550,11 @@ RouteEntryImplBase::RouteEntryImplBase(const VirtualHostImpl& vhost,
+               "not be stripped: {}",
+               path_redirect_);
+   }
+-  ENVOY_LOG(info, "route stats is {}, name is {}", route.stat_prefix(), route.name());
+   if (!route.stat_prefix().empty()) {
+     route_stats_context_ = std::make_unique<RouteStatsContext>(
+         factory_context.scope(), factory_context.routerContext().routeStatNames(), vhost.statName(),
+         route.stat_prefix());
+-  } else if (!route.name().empty()) {
+-    // Added by Ingress
+-    // use route_name as default stat_prefix
+-    route_stats_context_ = std::make_unique<RouteStatsContext>(
+-        factory_context.scope(), factory_context.routerContext().routeStatNames(), vhost.statName(),
+-        route.name());
+   }
+-  // End Added
+ }
+ 
+ bool RouteEntryImplBase::evaluateRuntimeMatch(const uint64_t random_value) const {
+@@ -1415,9 +1407,7 @@ VirtualHostImpl::VirtualHostImpl(
+       retry_shadow_buffer_limit_(PROTOBUF_GET_WRAPPED_OR_DEFAULT(
+           virtual_host, per_request_buffer_limit_bytes, std::numeric_limits<uint32_t>::max())),
+       include_attempt_count_in_request_(virtual_host.include_request_attempt_count()),
+-      include_attempt_count_in_response_(virtual_host.include_attempt_count_in_response()),
+-      virtual_cluster_catch_all_(*vcluster_scope_,
+-                                 factory_context.routerContext().virtualClusterStatNames()) {
+      include_attempt_count_in_response_(virtual_host.include_attempt_count_in_response()) {
+   switch (virtual_host.require_tls()) {
+   case envoy::config::route::v3::VirtualHost::NONE:
+     ssl_requirements_ = SslRequirements::None;
+@@ -1478,10 +1468,14 @@ VirtualHostImpl::VirtualHostImpl(
+     }
+   }
+ 
+-  for (const auto& virtual_cluster : virtual_host.virtual_clusters()) {
+-    virtual_clusters_.push_back(
+-        VirtualClusterEntry(virtual_cluster, *vcluster_scope_,
+-                            factory_context.routerContext().virtualClusterStatNames()));
+  if (!virtual_host.virtual_clusters().empty()) {
+    virtual_cluster_catch_all_ = std::make_unique<CatchAllVirtualCluster>(
+        *vcluster_scope_, factory_context.routerContext().virtualClusterStatNames());
+    for (const auto& virtual_cluster : virtual_host.virtual_clusters()) {
+      virtual_clusters_.push_back(
+          VirtualClusterEntry(virtual_cluster, *vcluster_scope_,
+                              factory_context.routerContext().virtualClusterStatNames()));
+    }
+   }
+ 
+   if (virtual_host.has_cors()) {
+@@ -1774,7 +1768,7 @@ VirtualHostImpl::virtualClusterFromEntries(const Http::HeaderMap& headers) const
+   }
+ 
+   if (!virtual_clusters_.empty()) {
+-    return &virtual_cluster_catch_all_;
+    return virtual_cluster_catch_all_.get();
+   }
+ 
+   return nullptr;
+diff --git a/source/common/router/config_impl.h b/source/common/router/config_impl.h
+index cf0ddf3..d83eb94 100644
+--- a/source/common/router/config_impl.h
+++ b/source/common/router/config_impl.h
+@@ -352,10 +352,10 @@ private:
+   const bool include_attempt_count_in_response_;
+   absl::optional<envoy::config::route::v3::RetryPolicy> retry_policy_;
+   absl::optional<envoy::config::route::v3::HedgePolicy> hedge_policy_;
+-  const CatchAllVirtualCluster virtual_cluster_catch_all_;
+ #if defined(ALIMESH)
+   std::vector<std::string> allow_server_names_;
+ #endif
+  std::unique_ptr<const CatchAllVirtualCluster> virtual_cluster_catch_all_;
+ };
+ 
+ using VirtualHostSharedPtr = std::shared_ptr<VirtualHostImpl>;
+diff --git a/source/common/router/scoped_config_impl.cc b/source/common/router/scoped_config_impl.cc
+index 594d571..6482615 100644
+--- a/source/common/router/scoped_config_impl.cc
+++ b/source/common/router/scoped_config_impl.cc
+@@ -7,6 +7,8 @@
+ #include "source/common/http/header_utility.h"
+ #endif
+ 
+#include "source/common/protobuf/utility.h"
+
+ namespace Envoy {
+ namespace Router {
+ 
+@@ -239,7 +241,8 @@ HeaderValueExtractorImpl::computeFragment(const Http::HeaderMap& headers) const
+ 
+ ScopedRouteInfo::ScopedRouteInfo(envoy::config::route::v3::ScopedRouteConfiguration&& config_proto,
+                                  ConfigConstSharedPtr&& route_config)
+-    : config_proto_(std::move(config_proto)), route_config_(std::move(route_config)) {
+    : config_proto_(std::move(config_proto)), route_config_(std::move(route_config)),
+      config_hash_(MessageUtil::hash(config_proto)) {
+   // TODO(stevenzzzz): Maybe worth a KeyBuilder abstraction when there are more than one type of
+   // Fragment.
+   for (const auto& fragment : config_proto_.key().fragments()) {
+diff --git a/source/common/router/scoped_config_impl.h b/source/common/router/scoped_config_impl.h
+index 9f6a1b2..28e2ee5 100644
+--- a/source/common/router/scoped_config_impl.h
+++ b/source/common/router/scoped_config_impl.h
+@@ -154,11 +154,13 @@ public:
+     return config_proto_;
+   }
+   const std::string& scopeName() const { return config_proto_.name(); }
+  uint64_t configHash() const { return config_hash_; }
+ 
+ private:
+   envoy::config::route::v3::ScopedRouteConfiguration config_proto_;
+   ScopeKey scope_key_;
+   ConfigConstSharedPtr route_config_;
+  const uint64_t config_hash_;
+ };
+ using ScopedRouteInfoConstSharedPtr = std::shared_ptr<const ScopedRouteInfo>;
+ // Ordered map for consistent config dumping.
+diff --git a/source/common/router/scoped_rds.cc b/source/common/router/scoped_rds.cc
+index 133e91e..9b2096e 100644
+--- a/source/common/router/scoped_rds.cc
+++ b/source/common/router/scoped_rds.cc
+@@ -245,6 +245,11 @@ bool ScopedRdsConfigSubscription::addOrUpdateScopes(
+         dynamic_cast<const envoy::config::route::v3::ScopedRouteConfiguration&>(
+             resource.get().resource());
+     const std::string scope_name = scoped_route_config.name();
+    if (const auto& scope_info_iter = scoped_route_map_.find(scope_name);
+        scope_info_iter != scoped_route_map_.end() &&
+        scope_info_iter->second->configHash() == MessageUtil::hash(scoped_route_config)) {
+      continue;
+    }
+     rds.set_route_config_name(scoped_route_config.route_configuration_name());
+     std::unique_ptr<RdsRouteConfigProviderHelper> rds_config_provider_helper;
+     std::shared_ptr<ScopedRouteInfo> scoped_route_info = nullptr;
+@@ -398,6 +403,7 @@ void ScopedRdsConfigSubscription::onRdsConfigUpdate(const std::string& scope_nam
+   auto new_scoped_route_info = std::make_shared<ScopedRouteInfo>(
+       envoy::config::route::v3::ScopedRouteConfiguration(iter->second->configProto()),
+       std::move(new_rds_config));
+  scoped_route_map_[new_scoped_route_info->scopeName()] = new_scoped_route_info;
+   applyConfigUpdate([new_scoped_route_info](ConfigProvider::ConfigConstSharedPtr config)
+                         -> ConfigProvider::ConfigConstSharedPtr {
+     auto* thread_local_scoped_config =
+diff --git a/source/common/router/scoped_rds.h b/source/common/router/scoped_rds.h
+index d21d812..a510c1f 100644
+--- a/source/common/router/scoped_rds.h
+++ b/source/common/router/scoped_rds.h
+@@ -104,7 +104,7 @@ struct ScopedRdsStats {
+ // A scoped RDS subscription to be used with the dynamic scoped RDS ConfigProvider.
+ class ScopedRdsConfigSubscription
+     : public Envoy::Config::DeltaConfigSubscriptionInstance,
+-      Envoy::Config::SubscriptionBase<envoy::config::route::v3::ScopedRouteConfiguration> {
+      public Envoy::Config::SubscriptionBase<envoy::config::route::v3::ScopedRouteConfiguration> {
+ public:
+   using ScopedRouteConfigurationMap =
+       std::map<std::string, envoy::config::route::v3::ScopedRouteConfiguration>;
+diff --git a/test/common/router/scoped_config_impl_test.cc b/test/common/router/scoped_config_impl_test.cc
+index f63f258..69a2f4b 100644
+--- a/test/common/router/scoped_config_impl_test.cc
+++ b/test/common/router/scoped_config_impl_test.cc
+@@ -452,6 +452,24 @@ TEST_F(ScopedRouteInfoTest, Creation) {
+   EXPECT_EQ(info_->scopeKey(), makeKey({"foo", "bar"}));
+ }
+ 
+// Tests that config hash changes if ScopedRouteConfiguration of the ScopedRouteInfo changes.
+TEST_F(ScopedRouteInfoTest, Hash) {
+  const envoy::config::route::v3::ScopedRouteConfiguration config_copy = scoped_route_config_;
+  info_ = std::make_unique<ScopedRouteInfo>(scoped_route_config_, route_config_);
+  EXPECT_EQ(info_->routeConfig().get(), route_config_.get());
+  EXPECT_TRUE(TestUtility::protoEqual(info_->configProto(), config_copy));
+  EXPECT_EQ(info_->scopeName(), "foo_scope");
+  EXPECT_EQ(info_->scopeKey(), makeKey({"foo", "bar"}));
+
+  const auto info2 = std::make_unique<ScopedRouteInfo>(scoped_route_config_, route_config_);
+  ASSERT_EQ(info2->configHash(), info_->configHash());
+
+  // Mutate the config and hash should be different now.
+  scoped_route_config_.set_on_demand(true);
+  const auto info3 = std::make_unique<ScopedRouteInfo>(scoped_route_config_, route_config_);
+  ASSERT_NE(info3->configHash(), info_->configHash());
+}
+
+ class ScopedConfigImplTest : public testing::Test {
+ public:
+   void SetUp() override {
+diff --git a/test/common/router/scoped_rds_test.cc b/test/common/router/scoped_rds_test.cc
+index 09b96a6..b4776c9 100644
+--- a/test/common/router/scoped_rds_test.cc
+++ b/test/common/router/scoped_rds_test.cc
+@@ -13,6 +13,7 @@
+ #include "envoy/stats/scope.h"
+ 
+ #include "source/common/config/api_version.h"
+#include "source/common/config/config_provider_impl.h"
+ #include "source/common/config/grpc_mux_impl.h"
+ #include "source/common/protobuf/message_validator_impl.h"
+ #include "source/common/router/scoped_rds.h"
+@@ -365,6 +366,48 @@ key:
+                             "Didn't find a registered implementation for name: 'filter.unknown'");
+ }
+ 
+// Test that scopes with same config as existing scopes will be skipped in a config push.
+TEST_F(ScopedRdsTest, UnchangedScopesAreSkipped) {
+  setup();
+  init_watcher_.expectReady();
+  const std::string config_yaml = R"EOF(
+name: foo_scope
+route_configuration_name: foo_routes
+key:
+  fragments:
+    - string_key: x-foo-key
+)EOF";
+  const auto resource = parseScopedRouteConfigurationFromYaml(config_yaml);
+  const std::string config_yaml2 = R"EOF(
+name: foo_scope2
+route_configuration_name: foo_routes
+key:
+  fragments:
+    - string_key: x-bar-key
+)EOF";
+  const auto resource_2 = parseScopedRouteConfigurationFromYaml(config_yaml2);
+
+  // Delta API.
+  const auto decoded_resources = TestUtility::decodeResources({resource, resource_2});
+  context_init_manager_.initialize(init_watcher_);
+  EXPECT_NO_THROW(srds_subscription_->onConfigUpdate(decoded_resources.refvec_, {}, "v1"));
+  EXPECT_EQ(1UL,
+            server_factory_context_.scope_.counter("foo.scoped_rds.foo_scoped_routes.config_reload")
+                .value());
+  EXPECT_EQ(2UL, all_scopes_.value());
+  pushRdsConfig({"foo_routes"}, "111");
+  Envoy::Router::ScopedRdsConfigSubscription* srds_delta_subscription =
+      static_cast<Envoy::Router::ScopedRdsConfigSubscription*>(srds_subscription_);
+  ASSERT_NE(srds_delta_subscription, nullptr);
+  ASSERT_EQ("v1", srds_delta_subscription->configInfo()->last_config_version_);
+  // Push again the same set of config with different version number, the config will be skipped.
+  EXPECT_NO_THROW(srds_subscription_->onConfigUpdate(decoded_resources.refvec_, {}, "123"));
+  ASSERT_EQ("v1", srds_delta_subscription->configInfo()->last_config_version_);
+  EXPECT_EQ(2UL,
+            server_factory_context_.scope_.counter("foo.scoped_rds.foo_scoped_routes.config_reload")
+                .value());
+}
+
+ // Test ignoring the optional unknown factory in the per-virtualhost typed config.
+ TEST_F(ScopedRdsTest, OptionalUnknownFactoryForPerVirtualHostTypedConfig) {
+   OptionalHttpFilters optional_http_filters;
--- a/envoy/1.20/patches/envoy/20240725-rename-original-host-header.patch
+++ b/envoy/1.20/patches/envoy/20240725-rename-original-host-header.patch
@@ -0,0 +1,13 @@
+diff --git a/source/common/http/headers.h b/source/common/http/headers.h
+index a7a8a3393e..6af4a2852d 100644
+--- a/source/common/http/headers.h
+++ b/source/common/http/headers.h
+@@ -123,7 +123,7 @@ public:
+     const LowerCaseString TriCostTime{"req-cost-time"};
+     const LowerCaseString TriStartTime{"req-start-time"};
+     const LowerCaseString TriRespStartTime{"resp-start-time"};
+-    const LowerCaseString EnvoyOriginalHost{"original-host"};
+    const LowerCaseString EnvoyOriginalHost{"x-envoy-original-host"};
+     const LowerCaseString HigressOriginalService{"x-higress-original-service"};
+   } AliExtendedValues;
+ #endif
--- a/envoy/1.20/patches/envoy/20240725-set-buffer-limit.patch
+++ b/envoy/1.20/patches/envoy/20240725-set-buffer-limit.patch
@@ -0,0 +1,43 @@
+diff --git a/source/extensions/common/wasm/context.cc b/source/extensions/common/wasm/context.cc
+index 9642d8abd3..410baa856f 100644
+--- a/source/extensions/common/wasm/context.cc
+++ b/source/extensions/common/wasm/context.cc
+@@ -62,6 +62,21 @@ constexpr absl::string_view CelStateKeyPrefix = "wasm.";
+ #if defined(ALIMESH)
+ constexpr std::string_view ClearRouteCacheKey = "clear_route_cache";
+ constexpr std::string_view DisableClearRouteCache = "off";
+constexpr std::string_view SetDecoderBufferLimit = "set_decoder_buffer_limit";
+constexpr std::string_view SetEncoderBufferLimit = "set_encoder_buffer_limit";
+
+bool stringViewToUint32(std::string_view str, uint32_t& out_value) {
+  try {
+    unsigned long temp = std::stoul(std::string(str));
+    if (temp <= std::numeric_limits<uint32_t>::max()) {
+      out_value = static_cast<uint32_t>(temp);
+      return true;
+    }
+  } catch (const std::exception& e) {
+    ENVOY_LOG_MISC(critical, "stringToUint exception '{}'", e.what());
+  }
+  return false;
+}
+ #endif
+ 
+ using HashPolicy = envoy::config::route::v3::RouteAction::HashPolicy;
+@@ -1280,6 +1295,16 @@ WasmResult Context::setProperty(std::string_view path, std::string_view value) {
+     } else {
+       disable_clear_route_cache_ = false;
+     }
+  } else if (path == SetDecoderBufferLimit && decoder_callbacks_) {
+    uint32_t buffer_limit;
+    if (stringViewToUint32(value, buffer_limit)) {
+      decoder_callbacks_->setDecoderBufferLimit(buffer_limit);
+    }
+  } else if (path == SetEncoderBufferLimit && encoder_callbacks_) {
+    uint32_t buffer_limit;
+    if (stringViewToUint32(value, buffer_limit)) {
+      encoder_callbacks_->setEncoderBufferLimit(buffer_limit);
+    }
+   }
+ #endif
+   if (!state->setValue(toAbslStringView(value))) {
--- a/envoy/1.20/patches/envoy/20240726-custom-span-tag.patch
+++ b/envoy/1.20/patches/envoy/20240726-custom-span-tag.patch
@@ -0,0 +1,106 @@
+diff --git a/envoy/stream_info/stream_info.h b/envoy/stream_info/stream_info.h
+index c6d82db4f4..09717673b0 100644
+--- a/envoy/stream_info/stream_info.h
+++ b/envoy/stream_info/stream_info.h
+@@ -613,7 +613,21 @@ public:
+    * @return the number of times the request was attempted upstream, absl::nullopt if the request
+    * was never attempted upstream.
+    */
+
+   virtual absl::optional<uint32_t> attemptCount() const PURE;
+
+#ifdef ALIMESH
+  /**
+   * @param key the filter state key set by wasm filter.
+   * @param value the filter state value set by wasm filter.
+   */
+  virtual void setCustomSpanTag(const std::string& key, const std::string& value) PURE;
+
+  /**
+   * @return the key-value map of filter states set by wasm filter.
+   */
+  virtual const std::unordered_map<std::string, std::string>& getCustomSpanTagMap() const PURE;
+#endif
+ };
+ 
+ } // namespace StreamInfo
+diff --git a/source/common/stream_info/stream_info_impl.h b/source/common/stream_info/stream_info_impl.h
+index 6ce2afe773..d5e7a80b37 100644
+--- a/source/common/stream_info/stream_info_impl.h
+++ b/source/common/stream_info/stream_info_impl.h
+@@ -291,6 +291,20 @@ struct StreamInfoImpl : public StreamInfo {
+ 
+   absl::optional<uint32_t> attemptCount() const override { return attempt_count_; }
+ 
+#ifdef ALIMESH
+  void setCustomSpanTag(const std::string& key, const std::string& value) override {
+    auto it = custom_span_tags_.find(key);
+    if (it != custom_span_tags_.end()) {
+      it->second = value;
+    } else {
+      custom_span_tags_.emplace(key, value);
+    }
+  }
+
+  const std::unordered_map<std::string, std::string>& getCustomSpanTagMap() const override {
+    return custom_span_tags_;
+  }
+#endif
+   TimeSource& time_source_;
+   const SystemTime start_time_;
+   const MonotonicTime start_time_monotonic_;
+@@ -350,6 +364,9 @@ private:
+   absl::optional<Upstream::ClusterInfoConstSharedPtr> upstream_cluster_info_;
+   std::string filter_chain_name_;
+   Tracing::Reason trace_reason_;
+#ifdef ALIMESH
+  std::unordered_map<std::string, std::string> custom_span_tags_;
+#endif
+ };
+ 
+ } // namespace StreamInfo
+diff --git a/source/common/tracing/http_tracer_impl.cc b/source/common/tracing/http_tracer_impl.cc
+index e55cf00e0a..f94e9101d7 100644
+--- a/source/common/tracing/http_tracer_impl.cc
+++ b/source/common/tracing/http_tracer_impl.cc
+@@ -214,6 +214,14 @@ void HttpTracerUtility::setCommonTags(Span& span, const Http::ResponseHeaderMap*
+ 
+   span.setTag(Tracing::Tags::get().Component, Tracing::Tags::get().Proxy);
+ 
+#ifdef ALIMESH
+  // Wasm filter state
+  const auto& custom_span_tags = stream_info.getCustomSpanTagMap();
+  for (const auto& it : custom_span_tags) {
+    span.setTag(it.first, it.second);
+  }
+#endif
+
+   if (nullptr != stream_info.upstreamHost()) {
+     span.setTag(Tracing::Tags::get().UpstreamCluster, stream_info.upstreamHost()->cluster().name());
+     span.setTag(Tracing::Tags::get().UpstreamClusterName,
+diff --git a/source/extensions/common/wasm/context.cc b/source/extensions/common/wasm/context.cc
+index 410baa856f..b11ecf1cd6 100644
+--- a/source/extensions/common/wasm/context.cc
+++ b/source/extensions/common/wasm/context.cc
+@@ -60,6 +60,7 @@ namespace {
+ constexpr absl::string_view CelStateKeyPrefix = "wasm.";
+ 
+ #if defined(ALIMESH)
+constexpr absl::string_view CustomeTraceSpanTagPrefix = "trace_span_tag.";
+ constexpr std::string_view ClearRouteCacheKey = "clear_route_cache";
+ constexpr std::string_view DisableClearRouteCache = "off";
+ constexpr std::string_view SetDecoderBufferLimit = "set_decoder_buffer_limit";
+@@ -1271,6 +1272,13 @@ WasmResult Context::setProperty(std::string_view path, std::string_view value) {
+   if (!stream_info) {
+     return WasmResult::NotFound;
+   }
+#ifdef ALIMESH
+  if (absl::StartsWith(absl::string_view{path.data(), path.size()}, CustomeTraceSpanTagPrefix)) {
+    stream_info->setCustomSpanTag(std::string(path.substr(CustomeTraceSpanTagPrefix.size())),
+                                  std::string(value));
+    return WasmResult::Ok;
+  }
+#endif
+   std::string key;
+   absl::StrAppend(&key, CelStateKeyPrefix, toAbslStringView(path));
+   CelState* state;
--- a/get_helm.sh
+++ b/get_helm.sh
@@ -0,0 +1,341 @@
+#!/usr/bin/env bash
+
+# Copyright The Helm Authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# The install script is based off of the MIT-licensed script from glide,
+# the package manager for Go: https://github.com/Masterminds/glide.sh/blob/master/get
+
+: ${BINARY_NAME:="helm"}
+: ${USE_SUDO:="true"}
+: ${DEBUG:="false"}
+: ${VERIFY_CHECKSUM:="true"}
+: ${VERIFY_SIGNATURES:="false"}
+: ${HELM_INSTALL_DIR:="/usr/local/bin"}
+: ${GPG_PUBRING:="pubring.kbx"}
+
+HAS_CURL="$(type "curl" &> /dev/null && echo true || echo false)"
+HAS_WGET="$(type "wget" &> /dev/null && echo true || echo false)"
+HAS_OPENSSL="$(type "openssl" &> /dev/null && echo true || echo false)"
+HAS_GPG="$(type "gpg" &> /dev/null && echo true || echo false)"
+HAS_GIT="$(type "git" &> /dev/null && echo true || echo false)"
+
+# initArch discovers the architecture for this system.
+initArch() {
+  ARCH=$(uname -m)
+  case $ARCH in
+    armv5*) ARCH="armv5";;
+    armv6*) ARCH="armv6";;
+    armv7*) ARCH="arm";;
+    aarch64) ARCH="arm64";;
+    x86) ARCH="386";;
+    x86_64) ARCH="amd64";;
+    i686) ARCH="386";;
+    i386) ARCH="386";;
+  esac
+}
+
+# initOS discovers the operating system for this system.
+initOS() {
+  OS=$(echo `uname`|tr '[:upper:]' '[:lower:]')
+
+  case "$OS" in
+    # Minimalist GNU for Windows
+    mingw*|cygwin*) OS='windows';;
+  esac
+}
+
+# runs the given command as root (detects if we are root already)
+runAsRoot() {
+  if [ $EUID -ne 0 -a "$USE_SUDO" = "true" ]; then
+    sudo "${@}"
+  else
+    "${@}"
+  fi
+}
+
+# verifySupported checks that the os/arch combination is supported for
+# binary builds, as well whether or not necessary tools are present.
+verifySupported() {
+  local supported="darwin-amd64\ndarwin-arm64\nlinux-386\nlinux-amd64\nlinux-arm\nlinux-arm64\nlinux-ppc64le\nlinux-s390x\nlinux-riscv64\nwindows-amd64\nwindows-arm64"
+  if ! echo "${supported}" | grep -q "${OS}-${ARCH}"; then
+    echo "No prebuilt binary for ${OS}-${ARCH}."
+    echo "To build from source, go to https://github.com/helm/helm"
+    exit 1
+  fi
+
+  if [ "${HAS_CURL}" != "true" ] && [ "${HAS_WGET}" != "true" ]; then
+    echo "Either curl or wget is required"
+    exit 1
+  fi
+
+  if [ "${VERIFY_CHECKSUM}" == "true" ] && [ "${HAS_OPENSSL}" != "true" ]; then
+    echo "In order to verify checksum, openssl must first be installed."
+    echo "Please install openssl or set VERIFY_CHECKSUM=false in your environment."
+    exit 1
+  fi
+
+  if [ "${VERIFY_SIGNATURES}" == "true" ]; then
+    if [ "${HAS_GPG}" != "true" ]; then
+      echo "In order to verify signatures, gpg must first be installed."
+      echo "Please install gpg or set VERIFY_SIGNATURES=false in your environment."
+      exit 1
+    fi
+    if [ "${OS}" != "linux" ]; then
+      echo "Signature verification is currently only supported on Linux."
+      echo "Please set VERIFY_SIGNATURES=false or verify the signatures manually."
+      exit 1
+    fi
+  fi
+
+  if [ "${HAS_GIT}" != "true" ]; then
+    echo "[WARNING] Could not find git. It is required for plugin installation."
+  fi
+}
+
+# checkDesiredVersion checks if the desired version is available.
+checkDesiredVersion() {
+  if [ "x$DESIRED_VERSION" == "x" ]; then
+    # Get tag from release URL
+    local latest_release_url="https://get.helm.sh/helm-latest-version"
+    local latest_release_response=""
+    if [ "${HAS_CURL}" == "true" ]; then
+      latest_release_response=$( curl -L --silent --show-error --fail "$latest_release_url" 2>&1 || true )
+    elif [ "${HAS_WGET}" == "true" ]; then
+      latest_release_response=$( wget "$latest_release_url" -q -O - 2>&1 || true )
+    fi
+    TAG=$( echo "$latest_release_response" | grep '^v[0-9]' )
+    if [ "x$TAG" == "x" ]; then
+      printf "Could not retrieve the latest release tag information from %s: %s\n" "${latest_release_url}" "${latest_release_response}"
+      exit 1
+    fi
+  else
+    TAG=$DESIRED_VERSION
+  fi
+}
+
+# checkHelmInstalledVersion checks which version of helm is installed and
+# if it needs to be changed.
+checkHelmInstalledVersion() {
+  if [[ -f "${HELM_INSTALL_DIR}/${BINARY_NAME}" ]]; then
+    local version=$("${HELM_INSTALL_DIR}/${BINARY_NAME}" version --template="{{ .Version }}")
+    if [[ "$version" == "$TAG" ]]; then
+      echo "Helm ${version} is already ${DESIRED_VERSION:-latest}"
+      return 0
+    else
+      echo "Helm ${TAG} is available. Changing from version ${version}."
+      return 1
+    fi
+  else
+    return 1
+  fi
+}
+
+# downloadFile downloads the latest binary package and also the checksum
+# for that binary.
+downloadFile() {
+  HELM_DIST="helm-$TAG-$OS-$ARCH.tar.gz"
+  DOWNLOAD_URL="https://get.helm.sh/$HELM_DIST"
+  CHECKSUM_URL="$DOWNLOAD_URL.sha256"
+  HELM_TMP_ROOT="$(mktemp -dt helm-installer-XXXXXX)"
+  HELM_TMP_FILE="$HELM_TMP_ROOT/$HELM_DIST"
+  HELM_SUM_FILE="$HELM_TMP_ROOT/$HELM_DIST.sha256"
+  echo "Downloading $DOWNLOAD_URL"
+  if [ "${HAS_CURL}" == "true" ]; then
+    curl -SsL "$CHECKSUM_URL" -o "$HELM_SUM_FILE"
+    curl -SsL "$DOWNLOAD_URL" -o "$HELM_TMP_FILE"
+  elif [ "${HAS_WGET}" == "true" ]; then
+    wget -q -O "$HELM_SUM_FILE" "$CHECKSUM_URL"
+    wget -q -O "$HELM_TMP_FILE" "$DOWNLOAD_URL"
+  fi
+}
+
+# verifyFile verifies the SHA256 checksum of the binary package
+# and the GPG signatures for both the package and checksum file
+# (depending on settings in environment).
+verifyFile() {
+  if [ "${VERIFY_CHECKSUM}" == "true" ]; then
+    verifyChecksum
+  fi
+  if [ "${VERIFY_SIGNATURES}" == "true" ]; then
+    verifySignatures
+  fi
+}
+
+# installFile installs the Helm binary.
+installFile() {
+  HELM_TMP="$HELM_TMP_ROOT/$BINARY_NAME"
+  mkdir -p "$HELM_TMP"
+  tar xf "$HELM_TMP_FILE" -C "$HELM_TMP"
+  HELM_TMP_BIN="$HELM_TMP/$OS-$ARCH/helm"
+  echo "Preparing to install $BINARY_NAME into ${HELM_INSTALL_DIR}"
+  runAsRoot cp "$HELM_TMP_BIN" "$HELM_INSTALL_DIR/$BINARY_NAME"
+  echo "$BINARY_NAME installed into $HELM_INSTALL_DIR/$BINARY_NAME"
+}
+
+# verifyChecksum verifies the SHA256 checksum of the binary package.
+verifyChecksum() {
+  printf "Verifying checksum... "
+  local sum=$(openssl sha1 -sha256 ${HELM_TMP_FILE} | awk '{print $2}')
+  local expected_sum=$(cat ${HELM_SUM_FILE})
+  if [ "$sum" != "$expected_sum" ]; then
+    echo "SHA sum of ${HELM_TMP_FILE} does not match. Aborting."
+    exit 1
+  fi
+  echo "Done."
+}
+
+# verifySignatures obtains the latest KEYS file from GitHub main branch
+# as well as the signature .asc files from the specific GitHub release,
+# then verifies that the release artifacts were signed by a maintainer's key.
+verifySignatures() {
+  printf "Verifying signatures... "
+  local keys_filename="KEYS"
+  local github_keys_url="https://raw.githubusercontent.com/helm/helm/main/${keys_filename}"
+  if [ "${HAS_CURL}" == "true" ]; then
+    curl -SsL "${github_keys_url}" -o "${HELM_TMP_ROOT}/${keys_filename}"
+  elif [ "${HAS_WGET}" == "true" ]; then
+    wget -q -O "${HELM_TMP_ROOT}/${keys_filename}" "${github_keys_url}"
+  fi
+  local gpg_keyring="${HELM_TMP_ROOT}/keyring.gpg"
+  local gpg_homedir="${HELM_TMP_ROOT}/gnupg"
+  mkdir -p -m 0700 "${gpg_homedir}"
+  local gpg_stderr_device="/dev/null"
+  if [ "${DEBUG}" == "true" ]; then
+    gpg_stderr_device="/dev/stderr"
+  fi
+  gpg --batch --quiet --homedir="${gpg_homedir}" --import "${HELM_TMP_ROOT}/${keys_filename}" 2> "${gpg_stderr_device}"
+  gpg --batch --no-default-keyring --keyring "${gpg_homedir}/${GPG_PUBRING}" --export > "${gpg_keyring}"
+  local github_release_url="https://github.com/helm/helm/releases/download/${TAG}"
+  if [ "${HAS_CURL}" == "true" ]; then
+    curl -SsL "${github_release_url}/helm-${TAG}-${OS}-${ARCH}.tar.gz.sha256.asc" -o "${HELM_TMP_ROOT}/helm-${TAG}-${OS}-${ARCH}.tar.gz.sha256.asc"
+    curl -SsL "${github_release_url}/helm-${TAG}-${OS}-${ARCH}.tar.gz.asc" -o "${HELM_TMP_ROOT}/helm-${TAG}-${OS}-${ARCH}.tar.gz.asc"
+  elif [ "${HAS_WGET}" == "true" ]; then
+    wget -q -O "${HELM_TMP_ROOT}/helm-${TAG}-${OS}-${ARCH}.tar.gz.sha256.asc" "${github_release_url}/helm-${TAG}-${OS}-${ARCH}.tar.gz.sha256.asc"
+    wget -q -O "${HELM_TMP_ROOT}/helm-${TAG}-${OS}-${ARCH}.tar.gz.asc" "${github_release_url}/helm-${TAG}-${OS}-${ARCH}.tar.gz.asc"
+  fi
+  local error_text="If you think this might be a potential security issue,"
+  error_text="${error_text}\nplease see here: https://github.com/helm/community/blob/master/SECURITY.md"
+  local num_goodlines_sha=$(gpg --verify --keyring="${gpg_keyring}" --status-fd=1 "${HELM_TMP_ROOT}/helm-${TAG}-${OS}-${ARCH}.tar.gz.sha256.asc" 2> "${gpg_stderr_device}" | grep -c -E '^\[GNUPG:\] (GOODSIG|VALIDSIG)')
+  if [[ ${num_goodlines_sha} -lt 2 ]]; then
+    echo "Unable to verify the signature of helm-${TAG}-${OS}-${ARCH}.tar.gz.sha256!"
+    echo -e "${error_text}"
+    exit 1
+  fi
+  local num_goodlines_tar=$(gpg --verify --keyring="${gpg_keyring}" --status-fd=1 "${HELM_TMP_ROOT}/helm-${TAG}-${OS}-${ARCH}.tar.gz.asc" 2> "${gpg_stderr_device}" | grep -c -E '^\[GNUPG:\] (GOODSIG|VALIDSIG)')
+  if [[ ${num_goodlines_tar} -lt 2 ]]; then
+    echo "Unable to verify the signature of helm-${TAG}-${OS}-${ARCH}.tar.gz!"
+    echo -e "${error_text}"
+    exit 1
+  fi
+  echo "Done."
+}
+
+# fail_trap is executed if an error occurs.
+fail_trap() {
+  result=$?
+  if [ "$result" != "0" ]; then
+    if [[ -n "$INPUT_ARGUMENTS" ]]; then
+      echo "Failed to install $BINARY_NAME with the arguments provided: $INPUT_ARGUMENTS"
+      help
+    else
+      echo "Failed to install $BINARY_NAME"
+    fi
+    echo -e "\tFor support, go to https://github.com/helm/helm."
+  fi
+  cleanup
+  exit $result
+}
+
+# testVersion tests the installed client to make sure it is working.
+testVersion() {
+  set +e
+  HELM="$(command -v $BINARY_NAME)"
+  if [ "$?" = "1" ]; then
+    echo "$BINARY_NAME not found. Is $HELM_INSTALL_DIR on your "'$PATH?'
+    exit 1
+  fi
+  set -e
+}
+
+# help provides possible cli installation arguments
+help () {
+  echo "Accepted cli arguments are:"
+  echo -e "\t[--help|-h ] ->> prints this help"
+  echo -e "\t[--version|-v <desired_version>] . When not defined it fetches the latest release from GitHub"
+  echo -e "\te.g. --version v3.0.0 or -v canary"
+  echo -e "\t[--no-sudo]  ->> install without sudo"
+}
+
+# cleanup temporary files to avoid https://github.com/helm/helm/issues/2977
+cleanup() {
+  if [[ -d "${HELM_TMP_ROOT:-}" ]]; then
+    rm -rf "$HELM_TMP_ROOT"
+  fi
+}
+
+# Execution
+
+#Stop execution on any error
+trap "fail_trap" EXIT
+set -e
+
+# Set debug if desired
+if [ "${DEBUG}" == "true" ]; then
+  set -x
+fi
+
+# Parsing input arguments (if any)
+export INPUT_ARGUMENTS="${@}"
+set -u
+while [[ $# -gt 0 ]]; do
+  case $1 in
+    '--version'|-v)
+       shift
+       if [[ $# -ne 0 ]]; then
+           export DESIRED_VERSION="${1}"
+           if [[ "$1" != "v"* ]]; then
+               echo "Expected version arg ('${DESIRED_VERSION}') to begin with 'v', fixing..."
+               export DESIRED_VERSION="v${1}"
+           fi
+       else
+           echo -e "Please provide the desired version. e.g. --version v3.0.0 or -v canary"
+           exit 0
+       fi
+       ;;
+    '--no-sudo')
+       USE_SUDO="false"
+       ;;
+    '--help'|-h)
+       help
+       exit 0
+       ;;
+    *) exit 1
+       ;;
+  esac
+  shift
+done
+set +u
+
+initArch
+initOS
+verifySupported
+checkDesiredVersion
+if ! checkHelmInstalledVersion; then
+  downloadFile
+  verifyFile
+  installFile
+fi
+testVersion
+cleanup
--- a/go.mod
+++ b/go.mod
@@ -255,7 +255,6 @@ require (
 	go.opentelemetry.io/proto/otlp v0.12.0 // indirect
 	go.starlark.net v0.0.0-20200306205701-8dd3e2ee1dd5 // indirect
 	go.uber.org/multierr v1.11.0 // indirect
-	go.uber.org/zap v1.24.0 // indirect
 	golang.org/x/crypto v0.17.0 // indirect
 	golang.org/x/exp v0.0.0-20230321023759-10a507213a29 // indirect
 	golang.org/x/mod v0.11.0 // indirect
@@ -304,7 +303,7 @@ replace istio.io/client-go => ./external/client-go

 replace istio.io/istio => ./external/istio

-replace github.com/caddyserver/certmagic => github.com/2456868764/certmagic v1.0.1
+replace github.com/caddyserver/certmagic => github.com/2456868764/certmagic v1.0.2

 require (
 	github.com/caddyserver/certmagic v0.20.0
@@ -313,6 +312,7 @@ require (
 	github.com/kylelemons/godebug v1.1.0
 	github.com/mholt/acmez v1.2.0
 	github.com/tidwall/gjson v1.17.0
+	go.uber.org/zap v1.24.0
 	golang.org/x/net v0.17.0
 	helm.sh/helm/v3 v3.7.1
 	k8s.io/apiextensions-apiserver v0.25.4
--- a/go.sum
+++ b/go.sum
@@ -61,8 +61,8 @@ dmitri.shuralyov.com/html/belt v0.0.0-20180602232347-f7d459c86be0/go.mod h1:JLBr
 dmitri.shuralyov.com/service/change v0.0.0-20181023043359-a85b471d5412/go.mod h1:a1inKt/atXimZ4Mv927x+r7UpyzRUf4emIoiiSC2TN4=
 dmitri.shuralyov.com/state v0.0.0-20180228185332-28bcc343414c/go.mod h1:0PRwlb0D6DFvNNtx+9ybjezNCa8XF0xaYcETyp6rHWU=
 git.apache.org/thrift.git v0.0.0-20180902110319-2566ecd5d999/go.mod h1:fPE2ZNJGynbRyZ4dJvy6G277gSllfV2HJqblrnkyeyg=
-github.com/2456868764/certmagic v1.0.1 h1:dRzow2Npe9llFTBhNVl0fVe8Yi/Q14ygNonlaZUyDZQ=
-github.com/2456868764/certmagic v1.0.1/go.mod h1:LOn81EQYMPajdew6Ln6SVdHPxPqPv6jwsUg92kiNlcQ=
+github.com/2456868764/certmagic v1.0.2 h1:xYoN4z6seONwT85llWXZcASvQME8TOSiSWQvLJsGGsE=
+github.com/2456868764/certmagic v1.0.2/go.mod h1:LOn81EQYMPajdew6Ln6SVdHPxPqPv6jwsUg92kiNlcQ=
 github.com/AdaLogics/go-fuzz-headers v0.0.0-20210929163055-e81b3f25be97/go.mod h1:WpB7kf89yJUETZxQnP1kgYPNwlT2jjdDYUCoxVggM3g=
 github.com/AlecAivazis/survey/v2 v2.3.6 h1:NvTuVHISgTHEHeBFqt6BHOe4Ny/NwGZr7w+F8S9ziyw=
 github.com/AlecAivazis/survey/v2 v2.3.6/go.mod h1:4AuI9b7RjAR+G7v9+C4YSlX/YL3K3cWNXgWXOhllqvI=
--- a/helm/core/Chart.yaml
+++ b/helm/core/Chart.yaml
@@ -1,5 +1,5 @@
 apiVersion: v2
-appVersion: 1.4.0
+appVersion: 1.4.2
 description: Helm chart for deploying higress gateways
 icon: https://higress.io/img/higress_logo_small.png
 home: http://higress.io/
@@ -10,4 +10,4 @@ name: higress-core
 sources:
 - http://github.com/alibaba/higress
 type: application
-version: 1.4.0
+version: 1.4.2
--- a/helm/core/templates/_helpers.tpl
+++ b/helm/core/templates/_helpers.tpl
@@ -97,7 +97,7 @@ higress: {{ include "controller.name" . }}
 {{- end }}

 {{- define "skywalking.enabled" -}}
-{{- if and .Values.skywalking.enabled .Values.skywalking.service.address }}
+{{- if and (hasKey .Values "tracing") .Values.tracing.enable (hasKey .Values.tracing "skywalking") .Values.tracing.skywalking.service }}
 true
 {{- end }}
 {{- end }}
--- a/helm/core/templates/configmap.yaml
+++ b/helm/core/templates/configmap.yaml
@@ -46,10 +46,6 @@
          address: {{ .Values.global.tracer.lightstep.address }}
          # Access Token used to communicate with the Satellite pool
          accessToken: {{ .Values.global.tracer.lightstep.accessToken }}
-      {{- else if eq .Values.global.proxy.tracer "zipkin" }}
-        zipkin:
-          # Address of the Zipkin collector
-          address: {{ .Values.global.tracer.zipkin.address | default (print "zipkin." .Release.Namespace ":9411") }}
      {{- else if eq .Values.global.proxy.tracer "datadog" }}
        datadog:
          # Address of the Datadog Agent
@@ -88,7 +84,7 @@
      {{- if .Values.global.enableHigressIstio }}
      discoveryAddress: {{ printf "istiod.%s.svc" .Values.global.istioNamespace }}:15012
      {{- else }}
-      discoveryAddress: higress-controller.{{.Release.Namespace}}.svc:15012
+      discoveryAddress: {{ include "controller.name" . }}.{{.Release.Namespace}}.svc:15012
      {{- end }}
      {{- end }}
      proxyStatsMatcher:
@@ -109,7 +105,17 @@ metadata:
  labels:
    {{- include "gateway.labels" . | nindent 4 }}    
 data:
-
+  higress: |-
+    {{- $existingConfig := lookup "v1" "ConfigMap" .Release.Namespace "higress-config" }}
+    {{- $existingData := dict }}
+    {{- if $existingConfig }}
+    {{- $existingData = index $existingConfig.data "higress" | default "{}" | fromYaml }}
+    {{- end }}
+    {{- $newData := dict }}
+    {{- if and (hasKey .Values "tracing") .Values.tracing.enable }}
+    {{- $_ := set $newData "tracing" .Values.tracing }}
+    {{- end }}
+    {{- toYaml (merge $existingData $newData) | nindent 4 }}
  # Configuration file for the mesh networks to be used by the Split Horizon EDS.
  meshNetworks: |-
  {{- if .Values.global.meshNetworks }}
@@ -170,8 +176,8 @@ data:
                      "endpoint": {
                        "address": {
                          "socket_address": {
-                            "address": "{{ .Values.skywalking.service.address }}",
-                            "port_value": "{{ .Values.skywalking.service.port }}"
+                            "address": "{{ .Values.tracing.skywalking.service }}",
+                            "port_value": "{{ .Values.tracing.skywalking.port }}"
                          }
                        }
                      }
--- a/helm/core/templates/controller-clusterrole.yaml
+++ b/helm/core/templates/controller-clusterrole.yaml
@@ -9,7 +9,7 @@ rules:
  # ingress controller
  - apiGroups: ["extensions", "networking.k8s.io"]
    resources: ["ingresses"]
-    verbs: ["get", "list", "watch"]
+    verbs: ["create", "get", "list", "watch", "update", "delete", "patch"]
  - apiGroups: ["extensions", "networking.k8s.io"]
    resources: ["ingresses/status"]
    verbs: ["*"]
@@ -36,7 +36,7 @@ rules:
  # Needed for multicluster secret reading, possibly ingress certs in the future
  - apiGroups: [""]
    resources: ["secrets"]
-    verbs: ["get", "watch", "list"]
+    verbs: ["get", "watch", "list", "create", "update", "delete", "patch"]

  - apiGroups: ["networking.higress.io"]
    resources: ["mcpbridges"]
@@ -61,12 +61,12 @@ rules:

  # discovery and routing
  - apiGroups: [""]
-    resources: ["pods", "nodes", "services", "namespaces", "endpoints"]
+    resources: ["pods", "nodes", "services", "namespaces", "endpoints", "deployments"]
    verbs: ["get", "list", "watch"]
  - apiGroups: ["discovery.k8s.io"]
    resources: ["endpointslices"]
    verbs: ["get", "list", "watch"]
-    
+
  # Istiod and bootstrap.
  - apiGroups: ["certificates.k8s.io"]
    resources:
@@ -100,7 +100,7 @@ rules:
  - apiGroups: ["multicluster.x-k8s.io"]
    resources: ["serviceimports"]
    verbs: ["get", "watch", "list"]
-    
+
  # sidecar injection controller
  - apiGroups: ["admissionregistration.k8s.io"]
    resources: ["mutatingwebhookconfigurations"]
--- a/helm/core/templates/controller-deployment.yaml
+++ b/helm/core/templates/controller-deployment.yaml
@@ -26,9 +26,70 @@ spec:
        {{- toYaml . | nindent 8 }}
      {{- end }}
      serviceAccountName: {{ include "controller.serviceAccountName" . }}
+      {{- if .Values.global.priorityClassName }}
+      priorityClassName: "{{ .Values.global.priorityClassName }}"
+      {{- end }}
      securityContext:
        {{- toYaml .Values.controller.podSecurityContext | nindent 8 }}
      containers:
+        - name: {{ .Chart.Name }}
+          securityContext:
+            {{- toYaml .Values.controller.securityContext | nindent 12 }}
+          image: "{{ .Values.controller.hub | default .Values.global.hub }}/{{ .Values.controller.image | default "higress" }}:{{ .Values.controller.tag | default .Chart.AppVersion }}"
+          args:
+          - "serve"
+          - --gatewaySelectorKey=higress
+          - --gatewaySelectorValue={{ .Release.Namespace }}-{{ include "gateway.name" . }}
+          - --gatewayHttpPort={{ .Values.gateway.httpPort }}
+          - --gatewayHttpsPort={{ .Values.gateway.httpsPort }}
+          {{- if not .Values.global.enableStatus }}
+          - --enableStatus={{ .Values.global.enableStatus }}
+          {{- end }}
+          - --ingressClass={{ .Values.global.ingressClass }}
+          {{- if .Values.global.watchNamespace }}
+          - --watchNamespace={{ .Values.global.watchNamespace }}
+          {{- end }}
+          - --enableAutomaticHttps={{ .Values.controller.automaticHttps.enabled }}
+          - --automaticHttpsEmail={{ .Values.controller.automaticHttps.email }}
+          env:
+          - name: POD_NAME
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: metadata.name
+          - name: POD_NAMESPACE
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: metadata.namespace
+          - name: SERVICE_ACCOUNT
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: spec.serviceAccountName
+          - name: DOMAIN_SUFFIX
+            value: {{ .Values.global.proxy.clusterDomain }}
+          {{- if .Values.controller.env }}
+          {{- range $key, $val := .Values.controller.env }}
+          - name: {{ $key }}
+            value: "{{ $val }}"
+          {{- end }}
+          {{- end }}
+          ports:
+            {{- range $idx, $port := .Values.controller.ports }}
+            - name: {{ $port.name }}
+              containerPort: {{ $port.port }}
+              protocol: {{ $port.protocol }}
+            {{- end }}
+          readinessProbe:
+            {{- toYaml .Values.controller.probe | nindent 12 }}
+          {{- if not (or .Values.global.local .Values.global.kind) }}
+          resources:
+            {{- toYaml .Values.controller.resources | nindent 12 }}
+          {{- end }}
+          volumeMounts:
+          - name: log
+            mountPath: /var/log
 {{- if not .Values.global.enableHigressIstio }}
        - name: discovery
          image: "{{ .Values.pilot.hub | default .Values.global.hub }}/{{ .Values.pilot.image | default "pilot" }}:{{ .Values.pilot.tag | default .Chart.AppVersion }}"
@@ -191,64 +252,6 @@ spec:
            mountPath: /cacerts
          {{- end }}
 {{- end }}
-        - name: {{ .Chart.Name }}
-          securityContext:
-            {{- toYaml .Values.controller.securityContext | nindent 12 }}
-          image: "{{ .Values.controller.hub | default .Values.global.hub }}/{{ .Values.controller.image | default "higress" }}:{{ .Values.controller.tag | default .Chart.AppVersion }}"
-          args:
-          - "serve"
-          - --gatewaySelectorKey=higress
-          - --gatewaySelectorValue={{ .Release.Namespace }}-{{ include "gateway.name" . }}
-          - --gatewayHttpPort={{ .Values.gateway.httpPort }}
-          - --gatewayHttpsPort={{ .Values.gateway.httpsPort }}
-          {{- if not .Values.global.enableStatus }}
-          - --enableStatus={{ .Values.global.enableStatus }}
-          {{- end }}
-          - --ingressClass={{ .Values.global.ingressClass }}
-          {{- if .Values.global.watchNamespace }}
-          - --watchNamespace={{ .Values.global.watchNamespace }}
-          {{- end }}
-          - --enableAutomaticHttps={{ .Values.controller.automaticHttps.enabled }}
-          - --automaticHttpsEmail={{ .Values.controller.automaticHttps.email }}
-          env:
-          - name: POD_NAME
-            valueFrom:
-              fieldRef:
-                apiVersion: v1
-                fieldPath: metadata.name
-          - name: POD_NAMESPACE
-            valueFrom:
-              fieldRef:
-                apiVersion: v1
-                fieldPath: metadata.namespace
-          - name: SERVICE_ACCOUNT
-            valueFrom:
-              fieldRef:
-                apiVersion: v1
-                fieldPath: spec.serviceAccountName
-          - name: DOMAIN_SUFFIX
-            value: {{ .Values.global.proxy.clusterDomain }}
-          {{- if .Values.controller.env }}
-          {{- range $key, $val := .Values.controller.env }}
-          - name: {{ $key }}
-            value: "{{ $val }}"
-          {{- end }}
-          {{- end }}
-          ports:
-            {{- range $idx, $port := .Values.controller.ports }}
-            - name: {{ $port.name }}
-              containerPort: {{ $port.port }}
-              protocol: {{ $port.protocol }}
-            {{- end }}
-          readinessProbe:
-            {{- toYaml .Values.controller.probe | nindent 12 }}
-          {{- if not (or .Values.global.local .Values.global.kind) }}
-          resources:
-            {{- toYaml .Values.controller.resources | nindent 12 }}
-          {{- end }}
-          volumeMounts:
-          - name: log
-            mountPath: /var/log
      {{- with .Values.controller.nodeSelector }}
      nodeSelector:
        {{- toYaml . | nindent 8 }}
--- a/helm/core/templates/daemonset.yaml
+++ b/helm/core/templates/daemonset.yaml
@@ -0,0 +1,332 @@
+{{- if eq .Values.gateway.kind "DaemonSet" -}}
+{{- $o11y := .Values.global.o11y  }}
+{{- $unprivilegedPortSupported := true }}
+{{- range $index, $node := (lookup "v1" "Node" "default" "").items }}
+    {{- $kernelVersion := $node.status.nodeInfo.kernelVersion }}
+    {{- if $kernelVersion }}
+      {{- $kernelVersion = regexFind "^(\\d+\\.\\d+\\.\\d+)" $kernelVersion }}
+      {{- if and $kernelVersion (semverCompare "<4.11.0" $kernelVersion) }}
+      {{-   $unprivilegedPortSupported = false }}
+      {{- end }}
+    {{- end }}
+{{- end -}}
+apiVersion: apps/v1
+kind: DaemonSet
+metadata:
+  name: {{ include "gateway.name" . }}
+  namespace: {{ .Release.Namespace }}
+  labels:
+    {{- include "gateway.labels" . | nindent 4}}
+  annotations:
+    {{- .Values.gateway.annotations | toYaml | nindent 4 }}
+spec:
+  selector:
+    matchLabels:
+      {{- include "gateway.selectorLabels" . | nindent 6 }}
+  template:
+    metadata:
+      annotations:
+      {{- if .Values.global.enableHigressIstio }}
+        "enableHigressIstio": "true"
+      {{- end }}
+      {{- if .Values.gateway.podAnnotations }}
+        {{- toYaml .Values.gateway.podAnnotations | nindent 8 }}
+      {{- end }}
+      labels:
+        sidecar.istio.io/inject: "false"
+        {{- with .Values.gateway.revision }}
+        istio.io/rev: {{ . }}
+        {{- end }}
+        {{- include "gateway.selectorLabels" . | nindent 8 }}
+    spec:
+      {{- with .Values.gateway.imagePullSecrets }}
+      imagePullSecrets:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
+      serviceAccountName: {{ include "gateway.serviceAccountName" . }}
+      {{- if .Values.global.priorityClassName }}
+      priorityClassName: "{{ .Values.global.priorityClassName }}"
+      {{- end }}
+      securityContext:
+      {{- if .Values.gateway.securityContext }}
+        {{- toYaml .Values.gateway.securityContext | nindent 8 }}
+      {{- else if and $unprivilegedPortSupported (and (not .Values.gateway.hostNetwork) (semverCompare ">=1.22-0" .Capabilities.KubeVersion.GitVersion)) }}
+        # Safe since 1.22: https://github.com/kubernetes/kubernetes/pull/103326
+        sysctls:
+        - name: net.ipv4.ip_unprivileged_port_start
+          value: "0"
+      {{- end }}
+      containers:
+        {{- if $o11y.enabled }}
+          {{- $config := $o11y.promtail }}
+        - name: promtail
+          image: {{ $config.image.repository }}:{{ $config.image.tag }}
+          imagePullPolicy: IfNotPresent
+          args:
+            - -config.file=/etc/promtail/promtail.yaml
+          env:
+            - name: 'HOSTNAME'
+              valueFrom:
+                fieldRef:
+                  fieldPath: 'spec.nodeName'
+          ports:
+            - containerPort: {{ $config.port }}
+              name: http-metrics
+              protocol: TCP
+          readinessProbe:
+            failureThreshold: 3
+            httpGet:
+              path: /ready
+              port: {{ $config.port }}
+              scheme: HTTP
+            initialDelaySeconds: 10
+            periodSeconds: 10
+            successThreshold: 1
+            timeoutSeconds: 1
+          volumeMounts:
+            - name: promtail-config
+              mountPath: "/etc/promtail"
+            - name: log
+              mountPath: /var/log/proxy
+            - name: tmp
+              mountPath: /tmp
+        {{- end }}
+        - name: higress-gateway
+          image: "{{ .Values.gateway.hub | default .Values.global.hub }}/{{ .Values.gateway.image | default "gateway" }}:{{ .Values.gateway.tag | default .Chart.AppVersion }}"
+          args:
+            - proxy
+            - router
+            - --domain
+            - $(POD_NAMESPACE).svc.cluster.local
+            - --proxyLogLevel=warning
+            - --proxyComponentLogLevel=misc:error
+            - --log_output_level=all:info
+            - --serviceCluster=higress-gateway
+          securityContext:
+          {{- if .Values.gateway.containerSecurityContext }}
+            {{- toYaml .Values.gateway.containerSecurityContext | nindent 12 }}
+          {{- else if and $unprivilegedPortSupported (and (not .Values.gateway.hostNetwork) (semverCompare ">=1.22-0" .Capabilities.KubeVersion.GitVersion)) }}
+            # Safe since 1.22: https://github.com/kubernetes/kubernetes/pull/103326
+            capabilities:
+              drop:
+              - ALL
+            allowPrivilegeEscalation: false
+            privileged: false
+          # When enabling lite metrics, the configuration template files need to be replaced.
+          {{- if not .Values.global.liteMetrics }}
+            readOnlyRootFilesystem: true
+          {{- end }}
+            runAsUser: 1337
+            runAsGroup: 1337
+            runAsNonRoot: true
+          {{- else }}
+            capabilities:
+              drop:
+              - ALL
+              add:
+              - NET_BIND_SERVICE
+            runAsUser: 0
+            runAsGroup: 1337
+            runAsNonRoot: false
+            allowPrivilegeEscalation: true
+          {{- end }}
+          env:
+          - name: NODE_NAME
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: spec.nodeName
+          - name: POD_NAME
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: metadata.name
+          - name: POD_NAMESPACE
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: metadata.namespace
+          - name: INSTANCE_IP
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: status.podIP
+          - name: HOST_IP
+            valueFrom:
+              fieldRef:
+                apiVersion: v1
+                fieldPath: status.hostIP
+          - name: SERVICE_ACCOUNT
+            valueFrom:
+              fieldRef:
+                fieldPath: spec.serviceAccountName
+          - name: PILOT_XDS_SEND_TIMEOUT
+            value: 60s
+          - name: PROXY_XDS_VIA_AGENT
+            value: "true"
+          - name: ENABLE_INGRESS_GATEWAY_SDS
+            value: "false"
+          - name: JWT_POLICY
+            value: {{ include "controller.jwtPolicy" . }}
+          - name: ISTIO_META_HTTP10
+            value: "1"
+          - name: ISTIO_META_CLUSTER_ID
+            value: "{{ $.Values.clusterName | default `Kubernetes` }}"
+          - name: INSTANCE_NAME
+            value: "higress-gateway"
+          {{- if .Values.global.liteMetrics }}
+          - name: LITE_METRICS
+            value: "on"
+          {{- end }}
+          {{- if include "skywalking.enabled" . }}
+          - name: ISTIO_BOOTSTRAP_OVERRIDE
+            value: /etc/istio/custom-bootstrap/custom_bootstrap.json
+          {{- end }}
+          {{- with .Values.gateway.networkGateway }}
+          - name: ISTIO_META_REQUESTED_NETWORK_VIEW
+            value: "{{.}}"
+          {{- end }}
+          {{- range $key, $val := .Values.env }}
+          - name: {{ $key }}
+            value: {{ $val | quote }}
+          {{- end }}
+          ports:
+          - containerPort: 15090
+            protocol: TCP
+            name: http-envoy-prom
+          {{- if or .Values.global.local .Values.global.kind }}
+          - containerPort: {{ .Values.gateway.httpPort }}
+            hostPort:  {{ .Values.gateway.httpPort }}
+            name: http
+            protocol: TCP
+          - containerPort:  {{ .Values.gateway.httpsPort }}
+            hostPort:  {{ .Values.gateway.httpsPort }}
+            name: https
+            protocol: TCP
+          {{- end }}
+          readinessProbe:
+            failureThreshold: {{ .Values.gateway.readinessFailureThreshold }}
+            httpGet:
+              path: /healthz/ready
+              port: 15021
+              scheme: HTTP
+            initialDelaySeconds: {{ .Values.gateway.readinessInitialDelaySeconds }}
+            periodSeconds: {{ .Values.gateway.readinessPeriodSeconds }}
+            successThreshold: {{ .Values.gateway.readinessSuccessThreshold }}
+            timeoutSeconds: {{ .Values.gateway.readinessTimeoutSeconds }}
+          {{- if not (or .Values.global.local .Values.global.kind) }}
+          resources:
+            {{- toYaml .Values.gateway.resources | nindent 12 }}
+          {{- end }}
+          volumeMounts:
+          {{- if eq (include "controller.jwtPolicy" .) "third-party-jwt" }}
+          - name: istio-token
+            mountPath: /var/run/secrets/tokens
+            readOnly: true
+          {{- end }}
+          - name: config
+            mountPath: /etc/istio/config
+          - name: istio-ca-root-cert
+            mountPath: /var/run/secrets/istio
+          - name: istio-data
+            mountPath: /var/lib/istio/data
+          - name: podinfo
+            mountPath: /etc/istio/pod
+          - name: proxy-socket
+            mountPath: /etc/istio/proxy
+          {{- if include "skywalking.enabled" . }}
+          - mountPath: /etc/istio/custom-bootstrap
+            name: custom-bootstrap-volume
+          {{- end }}
+          {{- if .Values.global.volumeWasmPlugins }}
+          - mountPath: /opt/plugins
+            name: local-wasmplugins-volume
+          {{- end }}
+          {{- if $o11y.enabled }}
+          - mountPath: /var/log/proxy
+            name: log
+          {{- end }}
+      {{- if .Values.gateway.hostNetwork }}
+      hostNetwork: {{ .Values.gateway.hostNetwork }}
+      dnsPolicy: ClusterFirstWithHostNet
+      {{- end }}
+      {{- with .Values.gateway.nodeSelector }}
+      nodeSelector:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
+      {{- with .Values.gateway.affinity }}
+      affinity:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
+      {{- with .Values.gateway.tolerations }}
+      tolerations:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
+      volumes:
+      {{- if eq (include "controller.jwtPolicy" .) "third-party-jwt" }}
+      - name: istio-token
+        projected:
+          sources:
+            - serviceAccountToken:
+                audience: istio-ca
+                expirationSeconds: 43200
+                path: istio-token
+      {{- end }}
+      - name: istio-ca-root-cert
+        configMap:
+      {{- if .Values.global.enableHigressIstio }}
+          name: istio-ca-root-cert
+      {{- else }}
+          name: higress-ca-root-cert
+      {{- end }}
+      - name: config
+        configMap:
+          name: higress-config
+      {{- if include "skywalking.enabled" . }}
+      - configMap:
+          defaultMode: 420
+          name: higress-custom-bootstrap
+        name: custom-bootstrap-volume
+      {{- end }}
+      - name: istio-data
+        emptyDir: {}
+      - name: proxy-socket
+        emptyDir: {}
+      {{- if $o11y.enabled }}
+      - name: log
+        emptyDir: {}
+      - name: tmp
+        emptyDir: {}
+      - name: promtail-config
+        configMap:
+          name: higress-promtail
+      {{- end }}
+      - name: podinfo
+        downwardAPI:
+          defaultMode: 420
+          items:
+          - fieldRef:
+              apiVersion: v1
+              fieldPath: metadata.labels
+            path: labels
+          - fieldRef:
+              apiVersion: v1
+              fieldPath: metadata.annotations
+            path: annotations
+          - path: cpu-request
+            resourceFieldRef:
+              containerName: higress-gateway
+              divisor: 1m
+              resource: requests.cpu
+          - path: cpu-limit
+            resourceFieldRef:
+              containerName: higress-gateway
+              divisor: 1m
+              resource: limits.cpu
+      {{- if .Values.global.volumeWasmPlugins }}
+      - name: local-wasmplugins-volume
+        hostPath:
+          path: /opt/plugins
+          type: Directory
+      {{- end }}
+{{- end }}
--- a/helm/core/templates/deployment.yaml
+++ b/helm/core/templates/deployment.yaml
@@ -1,3 +1,4 @@
+{{- if eq .Values.gateway.kind "Deployment" -}}
 {{- $o11y := .Values.global.o11y  }}
 {{- $unprivilegedPortSupported := true }}
 {{- range $index, $node := (lookup "v1" "Node" "default" "").items }}
@@ -58,6 +59,9 @@ spec:
        {{- toYaml . | nindent 8 }}
      {{- end }}
      serviceAccountName: {{ include "gateway.serviceAccountName" . }}
+      {{- if .Values.global.priorityClassName }}
+      priorityClassName: "{{ .Values.global.priorityClassName }}"
+      {{- end }}
      securityContext:
      {{- if .Values.gateway.securityContext }}
        {{- toYaml .Values.gateway.securityContext | nindent 8 }}
@@ -68,40 +72,6 @@ spec:
          value: "0"
      {{- end }}
      containers:
-        {{- if $o11y.enabled }}
-          {{- $config := $o11y.promtail }}
-        - name: promtail
-          image: {{ $config.image.repository }}:{{ $config.image.tag }}
-          imagePullPolicy: IfNotPresent
-          args:
-            - -config.file=/etc/promtail/promtail.yaml
-          env:
-            - name: 'HOSTNAME'
-              valueFrom:
-                fieldRef:
-                  fieldPath: 'spec.nodeName'
-          ports:
-            - containerPort: {{ $config.port }}
-              name: http-metrics
-              protocol: TCP
-          readinessProbe:
-            failureThreshold: 3
-            httpGet:
-              path: /ready
-              port: {{ $config.port }}
-              scheme: HTTP
-            initialDelaySeconds: 10
-            periodSeconds: 10
-            successThreshold: 1
-            timeoutSeconds: 1
-          volumeMounts:
-            - name: promtail-config
-              mountPath: "/etc/promtail"
-            - name: log
-              mountPath: /var/log/proxy
-            - name: tmp
-              mountPath: /tmp
-        {{- end }}
        - name: higress-gateway
          image: "{{ .Values.gateway.hub | default .Values.global.hub }}/{{ .Values.gateway.image | default "gateway" }}:{{ .Values.gateway.tag | default .Chart.AppVersion }}"
          args:
@@ -202,6 +172,9 @@ spec:
            value: {{ $val | quote }}
          {{- end }}
          ports:
+          - containerPort: 15020
+            protocol: TCP
+            name: istio-prom
          - containerPort: 15090
            protocol: TCP
            name: http-envoy-prom
@@ -241,7 +214,7 @@ spec:
            mountPath: /var/run/secrets/istio
          - name: istio-data
            mountPath: /var/lib/istio/data
-          - name: podinfo            
+          - name: podinfo
            mountPath: /etc/istio/pod
          - name: proxy-socket
            mountPath: /etc/istio/proxy
@@ -257,6 +230,40 @@ spec:
          - mountPath: /var/log/proxy
            name: log
          {{- end }}
+        {{- if $o11y.enabled }}
+          {{- $config := $o11y.promtail }}
+        - name: promtail
+          image: {{ $config.image.repository }}:{{ $config.image.tag }}
+          imagePullPolicy: IfNotPresent
+          args:
+            - -config.file=/etc/promtail/promtail.yaml
+          env:
+            - name: 'HOSTNAME'
+              valueFrom:
+                fieldRef:
+                  fieldPath: 'spec.nodeName'
+          ports:
+            - containerPort: {{ $config.port }}
+              name: http-metrics
+              protocol: TCP
+          readinessProbe:
+            failureThreshold: 3
+            httpGet:
+              path: /ready
+              port: {{ $config.port }}
+              scheme: HTTP
+            initialDelaySeconds: 10
+            periodSeconds: 10
+            successThreshold: 1
+            timeoutSeconds: 1
+          volumeMounts:
+            - name: promtail-config
+              mountPath: "/etc/promtail"
+            - name: log
+              mountPath: /var/log/proxy
+            - name: tmp
+              mountPath: /tmp
+        {{- end }}
      {{- if .Values.gateway.hostNetwork }}
      hostNetwork: {{ .Values.gateway.hostNetwork }}
      dnsPolicy: ClusterFirstWithHostNet
@@ -340,3 +347,4 @@ spec:
          path: /opt/plugins
          type: Directory
      {{- end }}
+{{- end }}
--- a/helm/core/templates/ingressclass.yaml
+++ b/helm/core/templates/ingressclass.yaml
@@ -0,0 +1,6 @@
+apiVersion: networking.k8s.io/v1
+kind: IngressClass
+metadata:
+  name: {{ .Values.global.ingressClass }}
+spec:
+  controller: higress.io/higress-controller
--- a/helm/core/templates/service.yaml
+++ b/helm/core/templates/service.yaml
@@ -15,6 +15,9 @@ spec:
 {{- with .Values.gateway.service.loadBalancerIP }}
  loadBalancerIP: "{{ . }}"
 {{- end }}
+{{- with .Values.gateway.service.loadBalancerClass }}
+  loadBalancerClass: "{{ . }}"
+{{- end }}
 {{- with .Values.gateway.service.loadBalancerSourceRanges }}
  loadBalancerSourceRanges:
 {{ toYaml . | indent 4 }}
--- a/helm/core/values.yaml
+++ b/helm/core/values.yaml
@@ -1,6 +1,6 @@
 revision: ""
 global:
-  liteMetrics: false
+  liteMetrics: true
  xdsMaxRecvMsgSize: "104857600"
  defaultUpstreamConcurrencyThreshold: 10000
  enableSRDS: true
@@ -178,9 +178,9 @@ global:
    # Default port for Pilot agent health checks. A value of 0 will disable health checking.
    statusPort: 15020

-    # Specify which tracer to use. One of: zipkin, lightstep, datadog, stackdriver.
+    # Specify which tracer to use. One of: lightstep, datadog, stackdriver.
    # If using stackdriver tracer outside GCP, set env GOOGLE_APPLICATION_CREDENTIALS to the GCP credential file.
-    tracer: "zipkin"
+    tracer: ""

    # Controls if sidecar is injected at the front of the container list and blocks the start of the other containers until the proxy is ready
    holdApplicationUntilProxyStarts: false
@@ -330,12 +330,8 @@ global:
      maxNumberOfAnnotations: 200
      # The global default max number of attributes per span.
      maxNumberOfAttributes: 200
-    zipkin:
-      # Host:Port for reporting trace data in zipkin format. If not specified, will default to
-      # zipkin service (port 9411) in the same namespace as the other istio components.
-      address: ""
-
  # Use the Mesh Control Protocol (MCP) for configuring Istiod. Requires an MCP source.
+
  useMCP: false

  # Observability (o11y) configurations
@@ -343,7 +339,7 @@ global:
    enabled: false
    promtail:
      image:
-        repository: grafana/promtail
+        repository: higress-registry.cn-hangzhou.cr.aliyuncs.com/higress/promtail
        tag: 2.9.4
      port: 3101
      resources:
@@ -396,6 +392,9 @@ gateway:
  replicas: 2
  image: gateway

+  # -- Use a `DaemonSet` or `Deployment`
+  kind: Deployment
+
  # The number of successive failed probes before indicating readiness failure.
  readinessFailureThreshold: 30

@@ -468,6 +467,7 @@ gateway:
      targetPort: 443
    annotations: {}
    loadBalancerIP: ""
+    loadBalancerClass: ""
    loadBalancerSourceRanges: []
    externalTrafficPolicy: ""

@@ -664,9 +664,15 @@ pilot:
  podLabels: {}


-# Skywalking config settings
-skywalking:
-  enabled: false
-  service:
-    address: ~
-    port: 11800
+# Tracing config settings
+tracing:
+  enable: false
+  sampling: 100
+  timeout: 500
+  skywalking:
+   # access_token: ""
+   service: ""
+   port: 11800
+  # zipkin:
+    # service: ""
+    # port: 9411
--- a/helm/higress/Chart.lock
+++ b/helm/higress/Chart.lock
@@ -1,9 +1,9 @@
 dependencies:
 - name: higress-core
  repository: file://../core
-  version: 1.4.0
+  version: 1.4.2
 - name: higress-console
  repository: https://higress.io/helm-charts/
-  version: 1.4.0
-digest: sha256:bf4c58ac28d4691907eab44a13eee398fc05ade95cdae07cb91d7e20ce4ba382
-generated: "2024-05-29T21:18:32.791995+08:00"
+  version: 1.4.2
+digest: sha256:31b557e55584e589b140ae9b89cfc8b99df91771c7d28465c3a2b06a4f35a192
+generated: "2024-07-26T13:53:23.225023+08:00"
--- a/helm/higress/Chart.yaml
+++ b/helm/higress/Chart.yaml
@@ -1,5 +1,5 @@
 apiVersion: v2
-appVersion: 1.4.0
+appVersion: 1.4.2
 description: Helm chart for deploying Higress gateways
 icon: https://higress.io/img/higress_logo_small.png
 home: http://higress.io/
@@ -12,9 +12,9 @@ sources:
 dependencies:
 - name: higress-core
  repository: "file://../core"
-  version: 1.4.0
+  version: 1.4.2
 - name: higress-console
  repository: "https://higress.io/helm-charts/"
-  version: 1.4.0
+  version: 1.4.2
 type: application
-version: 1.4.0
+version: 1.4.2
--- a/istio/1.12/patches/istio/20240607-fix-stats.patch
+++ b/istio/1.12/patches/istio/20240607-fix-stats.patch
@@ -0,0 +1,21 @@
+diff -Naur istio/tools/packaging/common/envoy_bootstrap.json istio-new/tools/packaging/common/envoy_bootstrap.json
+--- istio/tools/packaging/common/envoy_bootstrap.json	2024-06-07 16:50:21.000000000 +0800
+++ istio-new/tools/packaging/common/envoy_bootstrap.json	2024-06-07 16:47:42.000000000 +0800
+@@ -38,7 +38,7 @@
+     "stats_tags": [
+       {
+         "tag_name": "cluster_name",
+-        "regex": "^cluster\\.((.+?(\\..+?\\.svc\\.cluster\\.local)?)\\.)"
+        "regex": "^cluster\\.((.*?)\\.)(http1\\.|http2\\.|health_check\\.|zone\\.|external\\.|circuit_breakers\\.|[^\\.]+$)"
+       },
+       {
+         "tag_name": "tcp_prefix",
+@@ -58,7 +58,7 @@
+       },
+       {
+         "tag_name": "http_conn_manager_prefix",
+-        "regex": "^http\\.(((?:[_.[:digit:]]*|[_\\[\\]aAbBcCdDeEfF[:digit:]]*))\\.)"
+        "regex": "^http\\.(((outbound_([0-9]{1,3}\\.{0,1}){4}_\\d{0,5})|([^\\.]+))\\.)"
+       },
+       {
+         "tag_name": "listener_address",
--- a/istio/1.12/patches/istio/20240619-ai-stats.patch
+++ b/istio/1.12/patches/istio/20240619-ai-stats.patch
@@ -0,0 +1,53 @@
+diff -Naur istio/tools/packaging/common/envoy_bootstrap.json istio-new/tools/packaging/common/envoy_bootstrap.json
+--- istio/tools/packaging/common/envoy_bootstrap.json	2024-06-19 13:39:49.179159469 +0800
+++ istio-new/tools/packaging/common/envoy_bootstrap.json	2024-06-19 13:39:28.299159059 +0800
+@@ -37,6 +37,18 @@
+     "use_all_default_tags": false,
+     "stats_tags": [
+       {
+          "tag_name": "ai_route",
+          "regex": "^wasmcustom\\.route\\.((.*?)\\.)upstream"
+      },
+      {
+          "tag_name": "ai_cluster",
+          "regex": "^wasmcustom\\..*?\\.upstream\\.((.*?)\\.)model"
+      },
+      {
+          "tag_name": "ai_model",
+          "regex": "^wasmcustom\\..*?\\.model\\.((.*?)\\.)(input_token|output_token)"
+      },
+      {
+         "tag_name": "cluster_name",
+         "regex": "^cluster\\.((.*?)\\.)(http1\\.|http2\\.|health_check\\.|zone\\.|external\\.|circuit_breakers\\.|[^\\.]+$)"
+       },
+diff -Naur istio/tools/packaging/common/envoy_bootstrap_lite.json istio-new/tools/packaging/common/envoy_bootstrap_lite.json
+--- istio/tools/packaging/common/envoy_bootstrap_lite.json	2024-06-19 13:39:49.175159469 +0800
+++ istio-new/tools/packaging/common/envoy_bootstrap_lite.json	2024-06-19 13:38:52.283158352 +0800
+@@ -37,6 +37,18 @@
+     "use_all_default_tags": false,
+     "stats_tags": [
+       {
+          "tag_name": "ai_route",
+          "regex": "^wasmcustom\\.route\\.((.*?)\\.)upstream"
+      },
+      {
+          "tag_name": "ai_cluster",
+          "regex": "^wasmcustom\\..*?\\.upstream\\.((.*?)\\.)model"
+      },
+      {
+          "tag_name": "ai_model",
+          "regex": "^wasmcustom\\..*?\\.model\\.((.*?)\\.)(input_token|output_token)"
+      },
+      {
+         "tag_name": "response_code_class",
+         "regex": "_rq(_(\\dxx))$"
+       },
+@@ -60,7 +72,7 @@
+             "prefix": "vhost"
+           },
+           {
+-            "safe_regex": {"regex": "^http.*rds.*", "google_re2":{}}
+            "safe_regex": {"regex": "^http.*\\.rds\\..*", "google_re2":{}}
+           }
+         ]
+       }
--- a/pkg/bootstrap/server.go
+++ b/pkg/bootstrap/server.go
@@ -391,7 +391,7 @@ func (s *Server) initAutomaticHttps() error {
 		ServerAddress: s.CertHttpAddress,
 		Email:         s.AutomaticHttpsEmail,
 	}
-	certServer, err := cert.NewServer(s.kubeClient.Kube(), certOption)
+	certServer, err := cert.NewServer(s.kubeClient.Kube(), s.xdsServer, certOption)
 	if err != nil {
 		return err
 	}
--- a/pkg/cert/certmgr.go
+++ b/pkg/cert/certmgr.go
@@ -17,10 +17,15 @@ package cert
 import (
 	"context"
 	"fmt"
+	"os"
+	"reflect"
 	"sync"

 	"github.com/caddyserver/certmagic"
 	"github.com/mholt/acmez"
+	"go.uber.org/zap"
+	"go.uber.org/zap/zapcore"
+	"istio.io/istio/pilot/pkg/model"
 	"k8s.io/client-go/kubernetes"
 )

@@ -28,6 +33,10 @@ const (
 	EventCertObtained = "cert_obtained"
 )

+var (
+	cfg *certmagic.Config
+)
+
 type CertMgr struct {
 	cfg           *certmagic.Config
 	client        kubernetes.Interface
@@ -39,9 +48,10 @@ type CertMgr struct {
 	ingressSolver acmez.Solver
 	configMgr     *ConfigMgr
 	secretMgr     *SecretMgr
+	XDSUpdater    model.XDSUpdater
 }

-func InitCertMgr(opts *Option, clientSet kubernetes.Interface, config *Config) (*CertMgr, error) {
+func InitCertMgr(opts *Option, clientSet kubernetes.Interface, config *Config, XDSUpdater model.XDSUpdater, configMgr *ConfigMgr) (*CertMgr, error) {
 	CertLog.Infof("certmgr init config: %+v", config)
 	// Init certmagic config
 	// First make a pointer to a Cache as we need to reference the same Cache in
@@ -49,21 +59,29 @@ func InitCertMgr(opts *Option, clientSet kubernetes.Interface, config *Config) (
 	var cache *certmagic.Cache
 	var storage certmagic.Storage
 	storage, _ = NewConfigmapStorage(opts.Namespace, clientSet)
-	renewalWindowRatio := float64(config.RenewBeforeDays / RenewMaxDays)
+	renewalWindowRatio := float64(config.RenewBeforeDays) / float64(RenewMaxDays)
+	logger := zap.New(zapcore.NewCore(
+		zapcore.NewConsoleEncoder(zap.NewProductionEncoderConfig()),
+		os.Stderr,
+		zap.DebugLevel,
+	))
 	magicConfig := certmagic.Config{
 		RenewalWindowRatio: renewalWindowRatio,
 		Storage:            storage,
+		Logger:             logger,
 	}
 	cache = certmagic.NewCache(certmagic.CacheOptions{
 		GetConfigForCert: func(cert certmagic.Certificate) (*certmagic.Config, error) {
 			// Here we use New to get a valid Config associated with the same cache.
 			// The provided Config is used as a template and will be completed with
 			// any defaults that are set in the Default config.
-			return certmagic.New(cache, magicConfig), nil
+			return cfg, nil
 		},
+		Logger: logger,
 	})
 	// init certmagic
-	cfg := certmagic.New(cache, magicConfig)
+	cfg = certmagic.New(cache, magicConfig)
+
 	// Init certmagic acme
 	issuer := config.GetIssuer(IssuerTypeLetsencrypt)
 	if issuer == nil {
@@ -85,7 +103,6 @@ func InitCertMgr(opts *Option, clientSet kubernetes.Interface, config *Config) (
 	// init issuers
 	cfg.Issuers = []certmagic.Issuer{myACME}

-	configMgr, _ := NewConfigMgr(opts.Namespace, clientSet)
 	secretMgr, _ := NewSecretMgr(opts.Namespace, clientSet)

 	certMgr := &CertMgr{
@@ -97,6 +114,7 @@ func InitCertMgr(opts *Option, clientSet kubernetes.Interface, config *Config) (
 		configMgr:     configMgr,
 		secretMgr:     secretMgr,
 		cache:         cache,
+		XDSUpdater:    XDSUpdater,
 	}
 	certMgr.cfg.OnEvent = certMgr.OnEvent
 	return certMgr, nil
@@ -149,18 +167,31 @@ func (s *CertMgr) Reconcile(ctx context.Context, oldConfig *Config, newConfig *C
 		// sync email
 		s.myACME.Email = newIssuer.Email
 		// sync RenewalWindowRatio
-		s.cfg.RenewalWindowRatio = float64(newConfig.RenewBeforeDays / RenewMaxDays)
+		renewalWindowRatio := float64(newConfig.RenewBeforeDays) / float64(RenewMaxDays)
+		s.cfg.RenewalWindowRatio = renewalWindowRatio
 		// start cache
 		s.cache.Start()
 		// sync domains
-		s.manageSync(context.Background(), newDomains)
 		s.configMgr.SetConfig(newConfig)
+		CertLog.Infof("certMgr start to manageSync domains:+v%", newDomains)
+		s.manageSync(context.Background(), newDomains)
+		CertLog.Infof("certMgr manageSync domains done")
 	} else {
 		// stop cache  maintainAssets
 		s.cache.Stop()
 		s.configMgr.SetConfig(newConfig)
 	}

+	if oldConfig != nil && newConfig != nil {
+		if oldConfig.FallbackForInvalidSecret != newConfig.FallbackForInvalidSecret || !reflect.DeepEqual(oldConfig.CredentialConfig, newConfig.CredentialConfig) {
+			CertLog.Infof("ingress need to full push")
+			s.XDSUpdater.ConfigUpdate(&model.PushRequest{
+				Full:   true,
+				Reason: []model.TriggerReason{"higress-https-updated"},
+			})
+		}
+	}
+
 	return nil
 }

--- a/pkg/cert/config.go
+++ b/pkg/cert/config.go
@@ -86,22 +86,35 @@ func (c *Config) GetSecretNameByDomain(issuerName IssuerName, domain string) str
 	return ""
 }

+func ParseTLSSecret(tlsSecret string) (string, string) {
+	secrets := strings.Split(tlsSecret, "/")
+	switch len(secrets) {
+	case 1:
+		return "", tlsSecret
+	case 2:
+		return secrets[0], secrets[1]
+	}
+	return "", ""
+}
+
 func (c *Config) Validate() error {
 	// check acmeIssuer
-	if len(c.ACMEIssuer) == 0 {
-		return fmt.Errorf("acmeIssuer is empty")
-	}
-	for _, issuer := range c.ACMEIssuer {
-		switch issuer.Name {
-		case IssuerTypeLetsencrypt:
-			if issuer.Email == "" {
-				return fmt.Errorf("acmeIssuer %s email is empty", issuer.Name)
+	if c.AutomaticHttps {
+		if len(c.ACMEIssuer) == 0 {
+			return fmt.Errorf("no acmeIssuer configuration found when automaticHttps is enable")
+		}
+		for _, issuer := range c.ACMEIssuer {
+			switch issuer.Name {
+			case IssuerTypeLetsencrypt:
+				if issuer.Email == "" {
+					return fmt.Errorf("acmeIssuer %s email is empty", issuer.Name)
+				}
+				if !ValidateEmail(issuer.Email) {
+					return fmt.Errorf("acmeIssuer %s email %s is invalid", issuer.Name, issuer.Email)
+				}
+			default:
+				return fmt.Errorf("acmeIssuer name %s is not supported", issuer.Name)
 			}
-			if !ValidateEmail(issuer.Email) {
-				return fmt.Errorf("acmeIssuer %s email %s is invalid", issuer.Name, issuer.Email)
-			}
-		default:
-			return fmt.Errorf("acmeIssuer name %s is not supported", issuer.Name)
 		}
 	}
 	// check credentialConfig
@@ -111,14 +124,20 @@ func (c *Config) Validate() error {
 		}
 		if credential.TLSSecret == "" {
 			return fmt.Errorf("credentialConfig tlsSecret is empty")
+		} else {
+			ns, secret := ParseTLSSecret(credential.TLSSecret)
+			if ns == "" && secret == "" {
+				return fmt.Errorf("credentialConfig tlsSecret %s is not supported", credential.TLSSecret)
+			}
 		}
+
 		if credential.TLSIssuer == IssuerTypeLetsencrypt {
 			if len(credential.Domains) > 1 {
 				return fmt.Errorf("credentialConfig tlsIssuer %s only support one domain", credential.TLSIssuer)
 			}
 		}
 		if credential.TLSIssuer != IssuerTypeLetsencrypt && len(credential.TLSIssuer) > 0 {
-			return fmt.Errorf("credential tls issuer %s is not support", credential.TLSIssuer)
+			return fmt.Errorf("credential tls issuer %s is not supported", credential.TLSIssuer)
 		}
 	}

--- a/pkg/cert/config_test.go
+++ b/pkg/cert/config_test.go
@@ -120,3 +120,36 @@ func TestMatchSecretNameByDomain(t *testing.T) {
 		})
 	}
 }
+
+func TestParseTLSSecret(t *testing.T) {
+	tests := []struct {
+		tlsSecret          string
+		expectedNamespace  string
+		expectedSecretName string
+	}{
+		{
+			tlsSecret:          "example-com-tls",
+			expectedNamespace:  "",
+			expectedSecretName: "example-com-tls",
+		},
+
+		{
+			tlsSecret:          "kube-system/example-com-tls",
+			expectedNamespace:  "kube-system",
+			expectedSecretName: "example-com-tls",
+		},
+		{
+			tlsSecret:          "kube-system/example-com/wildcard",
+			expectedNamespace:  "",
+			expectedSecretName: "",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.tlsSecret, func(t *testing.T) {
+			resultNamespace, resultSecretName := ParseTLSSecret(tt.tlsSecret)
+			assert.Equal(t, tt.expectedNamespace, resultNamespace)
+			assert.Equal(t, tt.expectedSecretName, resultSecretName)
+		})
+	}
+}
--- a/pkg/cert/secret.go
+++ b/pkg/cert/secret.go
@@ -18,7 +18,6 @@ import (
 	"context"
 	"fmt"
 	"strconv"
-	"strings"
 	"time"

 	v1 "k8s.io/api/core/v1"
@@ -27,10 +26,6 @@ import (
 	"k8s.io/client-go/kubernetes"
 )

-const (
-	SecretNamePrefix = "higress-secret-"
-)
-
 type SecretMgr struct {
 	client    kubernetes.Interface
 	namespace string
@@ -46,13 +41,21 @@ func NewSecretMgr(namespace string, client kubernetes.Interface) (*SecretMgr, er
 }

 func (s *SecretMgr) Update(domain string, secretName string, privateKey []byte, certificate []byte, notBefore time.Time, notAfter time.Time, isRenew bool) error {
-	//secretName := s.getSecretName(domain)
-	secret := s.constructSecret(domain, privateKey, certificate, notBefore, notAfter, isRenew)
-	_, err := s.client.CoreV1().Secrets(s.namespace).Get(context.Background(), secretName, metav1.GetOptions{})
+	CertLog.Infof("update secret, domain:%s, secretName:%s, notBefore:%v, notAfter:%v, isRenew:%t", domain, secretName, notBefore, notAfter, isRenew)
+	name := secretName
+	namespace := s.namespace
+	namespaceP, secretP := ParseTLSSecret(secretName)
+	if namespaceP != "" {
+		namespace = namespaceP
+		name = secretP
+	}
+
+	secret := s.constructSecret(domain, name, namespace, privateKey, certificate, notBefore, notAfter, isRenew)
+	_, err := s.client.CoreV1().Secrets(namespace).Get(context.Background(), name, metav1.GetOptions{})
 	if err != nil {
 		if errors.IsNotFound(err) {
 			// create secret
-			_, err2 := s.client.CoreV1().Secrets(s.namespace).Create(context.Background(), secret, metav1.CreateOptions{})
+			_, err2 := s.client.CoreV1().Secrets(namespace).Create(context.Background(), secret, metav1.CreateOptions{})
 			return err2
 		}
 		return err
@@ -61,7 +64,7 @@ func (s *SecretMgr) Update(domain string, secretName string, privateKey []byte,
 	if _, ok := secret.Annotations["higress.io/cert-domain"]; !ok {
 		return fmt.Errorf("the secret name %s is not automatic https secret name for the domain:%s, please rename it in config", secretName, domain)
 	}
-	_, err1 := s.client.CoreV1().Secrets(s.namespace).Update(context.Background(), secret, metav1.UpdateOptions{})
+	_, err1 := s.client.CoreV1().Secrets(namespace).Update(context.Background(), secret, metav1.UpdateOptions{})
 	if err1 != nil {
 		return err1
 	}
@@ -69,23 +72,13 @@ func (s *SecretMgr) Update(domain string, secretName string, privateKey []byte,
 	return nil
 }

-func (s *SecretMgr) Delete(domain string) error {
-	secretName := s.getSecretName(domain)
-	err := s.client.CoreV1().Secrets(s.namespace).Delete(context.Background(), secretName, metav1.DeleteOptions{})
-	return err
-}
-
-func (s *SecretMgr) getSecretName(domain string) string {
-	return SecretNamePrefix + strings.ReplaceAll(strings.TrimSpace(domain), ".", "-")
-}
-
-func (s *SecretMgr) constructSecret(domain string, privateKey []byte, certificate []byte, notBefore time.Time, notAfter time.Time, isRenew bool) *v1.Secret {
-	secretName := s.getSecretName(domain)
+func (s *SecretMgr) constructSecret(domain string, name string, namespace string, privateKey []byte, certificate []byte, notBefore time.Time, notAfter time.Time, isRenew bool) *v1.Secret {
 	annotationMap := make(map[string]string, 0)
 	annotationMap["higress.io/cert-domain"] = domain
 	annotationMap["higress.io/cert-notAfter"] = notAfter.Format("2006-01-02 15:04:05")
 	annotationMap["higress.io/cert-notBefore"] = notBefore.Format("2006-01-02 15:04:05")
 	annotationMap["higress.io/cert-renew"] = strconv.FormatBool(isRenew)
+	annotationMap["higress.io/cert-source"] = string(IssuerTypeLetsencrypt)
 	if isRenew {
 		annotationMap["higress.io/cert-renew-time"] = time.Now().Format("2006-01-02 15:04:05")
 	}
@@ -97,8 +90,8 @@ func (s *SecretMgr) constructSecret(domain string, privateKey []byte, certificat
 	dataMap["tls.crt"] = certificate
 	secret := &v1.Secret{
 		ObjectMeta: metav1.ObjectMeta{
-			Name:        secretName,
-			Namespace:   s.namespace,
+			Name:        name,
+			Namespace:   namespace,
 			Annotations: annotationMap,
 		},
 		Type: v1.SecretTypeTLS,
--- a/pkg/cert/server.go
+++ b/pkg/cert/server.go
@@ -22,6 +22,7 @@ import (
 	"time"

 	"github.com/caddyserver/certmagic"
+	"istio.io/istio/pilot/pkg/model"
 	"k8s.io/client-go/kubernetes"
 )

@@ -37,12 +38,14 @@ type Server struct {
 	clientSet  kubernetes.Interface
 	controller *Controller
 	certMgr    *CertMgr
+	XDSUpdater model.XDSUpdater
 }

-func NewServer(clientSet kubernetes.Interface, opts *Option) (*Server, error) {
+func NewServer(clientSet kubernetes.Interface, XDSUpdater model.XDSUpdater, opts *Option) (*Server, error) {
 	server := &Server{
-		clientSet: clientSet,
-		opts:      opts,
+		clientSet:  clientSet,
+		opts:       opts,
+		XDSUpdater: XDSUpdater,
 	}
 	return server, nil
 }
@@ -65,7 +68,7 @@ func (s *Server) InitServer() error {
 		return err
 	}
 	// init certmgr
-	certMgr, err := InitCertMgr(s.opts, s.clientSet, defaultConfig) // config and start
+	certMgr, err := InitCertMgr(s.opts, s.clientSet, defaultConfig, s.XDSUpdater, configMgr) // config and start
 	s.certMgr = certMgr
 	// init controller
 	controller, err := NewController(s.clientSet, s.opts.Namespace, certMgr, configMgr)
--- a/pkg/cert/storage.go
+++ b/pkg/cert/storage.go
@@ -32,7 +32,7 @@ import (
 )

 const (
-	CertificatesPrefix              = "/certificates"
+	CertificatesPrefix              = "certificates"
 	ConfigmapStoreCertficatesPrefix = "higress-cert-store-certificates-"
 	ConfigmapStoreDefaultName       = "higress-cert-store-default"
 )
@@ -155,7 +155,7 @@ func (s *ConfigmapStorage) List(ctx context.Context, prefix string, recursive bo
 	// Check if the prefix corresponds to a specific key
 	hashPrefix := fastHash([]byte(prefix))
 	if strings.HasPrefix(prefix, CertificatesPrefix) {
-		// If the prefix is "/certificates", get all ConfigMaps and traverse each one
+		// If the prefix is "certificates/", get all ConfigMaps and traverse each one
 		// List all ConfigMaps in the namespace with label higress.io/cert-https=true
 		configmaps, err := s.client.CoreV1().ConfigMaps(s.namespace).List(ctx, metav1.ListOptions{FieldSelector: "metadata.annotations['higress.io/cert-https'] == 'true'"})
 		if err != nil {
@@ -289,14 +289,29 @@ func (s *ConfigmapStorage) String() string {
 	return "ConfigmapStorage"
 }

+// getConfigmapStoreNameByKey determines the storage name for a given key.
+// It checks if the key starts with 'certificates/' and if so, the key pattern should match one of the following:
+// 'certificates/<issuerKey>/<domain>/<domain>.json',
+// 'certificates/<issuerKey>/<domain>/<domain>.crt',
+// or 'certificates/<issuerKey>/<domain>/<domain>.key'.
+// It then returns the corresponding ConfigMap name.
+// If the key does not start with 'certificates/', it returns the default store name.
+//
+// Parameters:
+//
+// key - The configuration map key that needs to be mapped to a storage name.
+//
+// Returns:
+//
+// string - The calculated or default storage name based on the key.
 func (s *ConfigmapStorage) getConfigmapStoreNameByKey(key string) string {
-	parts := strings.SplitN(key, "/", 10)
-	if len(parts) >= 4 && parts[1] == "certificates" {
-		domain := strings.TrimSuffix(parts[3], ".crt")
-		domain = strings.TrimSuffix(domain, ".key")
-		domain = strings.TrimSuffix(domain, ".json")
-		issuerKey := parts[2]
-		return ConfigmapStoreCertficatesPrefix + fastHash([]byte(issuerKey+domain))
+	if strings.HasPrefix(key, "certificates/") {
+		parts := strings.Split(key, "/")
+		if len(parts) >= 4 && parts[0] == "certificates" {
+			domain := parts[2]
+			issuerKey := parts[1]
+			return ConfigmapStoreCertficatesPrefix + fastHash([]byte(issuerKey+domain))
+		}
 	}
 	return ConfigmapStoreDefaultName
 }
--- a/pkg/cert/storage_test.go
+++ b/pkg/cert/storage_test.go
@@ -39,22 +39,29 @@ func TestGetConfigmapStoreNameByKey(t *testing.T) {
 	}{
 		{
 			name:     "certificate crt",
-			key:      "/certificates/issuerKey/domain.crt",
+			key:      "certificates/issuerKey/domain/domain.crt",
 			expected: "higress-cert-store-certificates-" + fastHash([]byte("issuerKey"+"domain")),
 		},
+
+		{
+			name:     "47.237.14.136.sslip.io crt",
+			key:      "certificates/acme-v02.api.letsencrypt.org-directory/47.237.14.136.sslip.io/47.237.14.136.sslip.io.crt",
+			expected: "higress-cert-store-certificates-" + fastHash([]byte("acme-v02.api.letsencrypt.org-directory"+"47.237.14.136.sslip.io")),
+		},
+
 		{
 			name:     "certificate meta",
-			key:      "/certificates/issuerKey/domain.json",
+			key:      "certificates/issuerKey/domain/domain.json",
 			expected: "higress-cert-store-certificates-" + fastHash([]byte("issuerKey"+"domain")),
 		},
 		{
 			name:     "certificate key",
-			key:      "/certificates/issuerKey/domain.key",
+			key:      "certificates/issuerKey/domain/domain.key",
 			expected: "higress-cert-store-certificates-" + fastHash([]byte("issuerKey"+"domain")),
 		},
 		{
 			name:     "user key",
-			key:      "/users/hello/2",
+			key:      "users/hello/2",
 			expected: "higress-cert-store-default",
 		},
 		{
@@ -82,7 +89,7 @@ func TestExists(t *testing.T) {
 	assert.NoError(t, err)

 	// Store a test key
-	testKey := "/certificates/issuer1/domain1.crt"
+	testKey := "certificates/issuer1/domain1/domain1.crt"
 	err = storage.Store(context.Background(), testKey, []byte("test-data"))
 	assert.NoError(t, err)

@@ -94,17 +101,17 @@ func TestExists(t *testing.T) {
 	}{
 		{
 			name:        "Existing Key",
-			key:         "/certificates/issuer1/domain1.crt",
+			key:         "certificates/issuer1/domain1/domain1.crt",
 			shouldExist: true,
 		},
 		{
 			name:        "Non-Existent Key1",
-			key:         "/certificates/issuer2/domain2.crt",
+			key:         "certificates/issuer2/domain2/domain2.crt",
 			shouldExist: false,
 		},
 		{
 			name:        "Non-Existent Key2",
-			key:         "/users/hello/a",
+			key:         "users/hello/a",
 			shouldExist: false,
 		},
 		// Add more test cases as needed
@@ -129,7 +136,7 @@ func TestLoad(t *testing.T) {
 	assert.NoError(t, err)

 	// Store a test key
-	testKey := "/certificates/issuer1/domain1.crt"
+	testKey := "certificates/issuer1/domain1/domain1.crt"
 	testValue := []byte("test-data")
 	err = storage.Store(context.Background(), testKey, testValue)
 	assert.NoError(t, err)
@@ -143,13 +150,13 @@ func TestLoad(t *testing.T) {
 	}{
 		{
 			name:        "Existing Key",
-			key:         "/certificates/issuer1/domain1.crt",
+			key:         "certificates/issuer1/domain1/domain1.crt",
 			expected:    testValue,
 			shouldError: false,
 		},
 		{
 			name:        "Non-Existent Key",
-			key:         "/certificates/issuer2/domain2.crt",
+			key:         "certificates/issuer2/domain2/domain2.crt",
 			expected:    nil,
 			shouldError: true,
 		},
@@ -192,28 +199,28 @@ func TestStore(t *testing.T) {
 		shouldError           bool
 	}{
 		{
-			name:                  "Store Key with /certificates prefix",
-			key:                   "/certificates/issuer1/domain1.crt",
+			name:                  "Store Key with certificates prefix",
+			key:                   "certificates/issuer1/domain1/domain1.crt",
 			value:                 []byte("test-data1"),
-			expected:              map[string]string{fastHash([]byte("/certificates/issuer1/domain1.crt")): `{"k":"/certificates/issuer1/domain1.crt","v":"dGVzdC1kYXRhMQ=="}`},
+			expected:              map[string]string{fastHash([]byte("certificates/issuer1/domain1/domain1.crt")): `{"k":"certificates/issuer1/domain1/domain1.crt","v":"dGVzdC1kYXRhMQ=="}`},
 			expectedConfigmapName: "higress-cert-store-certificates-" + fastHash([]byte("issuer1"+"domain1")),
 			shouldError:           false,
 		},
 		{
-			name:  "Store Key with /certificates prefix (additional data)",
-			key:   "/certificates/issuer2/domain2.crt",
+			name:  "Store Key with certificates prefix (additional data)",
+			key:   "certificates/issuer2/domain2/domain2.crt",
 			value: []byte("test-data2"),
 			expected: map[string]string{
-				fastHash([]byte("/certificates/issuer2/domain2.crt")): `{"k":"/certificates/issuer2/domain2.crt","v":"dGVzdC1kYXRhMg=="}`,
+				fastHash([]byte("certificates/issuer2/domain2/domain2.crt")): `{"k":"certificates/issuer2/domain2/domain2.crt","v":"dGVzdC1kYXRhMg=="}`,
 			},
 			expectedConfigmapName: "higress-cert-store-certificates-" + fastHash([]byte("issuer2"+"domain2")),
 			shouldError:           false,
 		},
 		{
-			name:                  "Store Key without /certificates prefix",
-			key:                   "/other/path/data.txt",
+			name:                  "Store Key without certificates prefix",
+			key:                   "other/path/data.txt",
 			value:                 []byte("test-data3"),
-			expected:              map[string]string{fastHash([]byte("/other/path/data.txt")): `{"k":"/other/path/data.txt","v":"dGVzdC1kYXRhMw=="}`},
+			expected:              map[string]string{fastHash([]byte("other/path/data.txt")): `{"k":"other/path/data.txt","v":"dGVzdC1kYXRhMw=="}`},
 			expectedConfigmapName: "higress-cert-store-default",
 			shouldError:           false,
 		},
@@ -256,17 +263,17 @@ func TestList(t *testing.T) {
 	// Store some test data
 	// Store some test data
 	testKeys := []string{
-		"/certificates/issuer1/domain1.crt",
-		"/certificates/issuer1/domain2.crt",
-		"/certificates/issuer1/domain3.crt", // Added another domain for issuer1
-		"/certificates/issuer2/domain4.crt",
-		"/certificates/issuer2/domain5.crt",
-		"/certificates/issuer3/subdomain1/domain6.crt",            // Two-level subdirectory under issuer3
-		"/certificates/issuer3/subdomain1/subdomain2/domain7.crt", // Two more levels under issuer3
-		"/other-prefix/key1/file1",
-		"/other-prefix/key1/file2",
-		"/other-prefix/key2/file3",
-		"/other-prefix/key2/file4",
+		"certificates/issuer1/domain1/domain1.crt",
+		"certificates/issuer1/domain2/domain2.crt",
+		"certificates/issuer1/domain3/domain3.crt", // Added another domain for issuer1
+		"certificates/issuer2/domain4/domain4.crt",
+		"certificates/issuer2/domain5/domain5.crt",
+		"certificates/issuer3/domain6/domain6.crt",               // Two-level subdirectory under issuer3
+		"certificates/issuer3/subdomain1/subdomain2/domain7.crt", // Two more levels under issuer3
+		"other-prefix/key1/file1",
+		"other-prefix/key1/file2",
+		"other-prefix/key2/file3",
+		"other-prefix/key2/file4",
 	}

 	for _, key := range testKeys {
@@ -283,34 +290,34 @@ func TestList(t *testing.T) {
 	}{
 		{
 			name:      "List Certificates (Non-Recursive)",
-			prefix:    "/certificates",
+			prefix:    "certificates",
 			recursive: false,
-			expected:  []string{"/certificates/issuer1", "/certificates/issuer2", "/certificates/issuer3"},
+			expected:  []string{"certificates/issuer1", "certificates/issuer2", "certificates/issuer3"},
 		},
 		{
 			name:      "List Certificates (Recursive)",
-			prefix:    "/certificates",
+			prefix:    "certificates",
 			recursive: true,
-			expected:  []string{"/certificates/issuer1/domain1.crt", "/certificates/issuer1/domain2.crt", "/certificates/issuer1/domain3.crt", "/certificates/issuer2/domain4.crt", "/certificates/issuer2/domain5.crt", "/certificates/issuer3/subdomain1/domain6.crt", "/certificates/issuer3/subdomain1/subdomain2/domain7.crt"},
+			expected:  []string{"certificates/issuer1/domain1/domain1.crt", "certificates/issuer1/domain2/domain2.crt", "certificates/issuer1/domain3/domain3.crt", "certificates/issuer2/domain4/domain4.crt", "certificates/issuer2/domain5/domain5.crt", "certificates/issuer3/domain6/domain6.crt", "certificates/issuer3/subdomain1/subdomain2/domain7.crt"},
 		},
 		{
 			name:      "List Other Prefix (Non-Recursive)",
-			prefix:    "/other-prefix",
+			prefix:    "other-prefix",
 			recursive: false,
-			expected:  []string{"/other-prefix/key1", "/other-prefix/key2"},
+			expected:  []string{"other-prefix/key1", "other-prefix/key2"},
 		},

 		{
 			name:      "List Other Prefix (Non-Recursive)",
-			prefix:    "/other-prefix/key1",
+			prefix:    "other-prefix/key1",
 			recursive: false,
-			expected:  []string{"/other-prefix/key1/file1", "/other-prefix/key1/file2"},
+			expected:  []string{"other-prefix/key1/file1", "other-prefix/key1/file2"},
 		},
 		{
 			name:      "List Other Prefix (Recursive)",
-			prefix:    "/other-prefix",
+			prefix:    "other-prefix",
 			recursive: true,
-			expected:  []string{"/other-prefix/key1/file1", "/other-prefix/key1/file2", "/other-prefix/key2/file3", "/other-prefix/key2/file4"},
+			expected:  []string{"other-prefix/key1/file1", "other-prefix/key1/file2", "other-prefix/key2/file3", "other-prefix/key2/file4"},
 		},
 	}

--- a/pkg/cmd/hgctl/common.go
+++ b/pkg/cmd/hgctl/common.go
@@ -15,7 +15,8 @@
 package hgctl

 const (
-	yamlOutput  = "yaml"
-	jsonOutput  = "json"
-	flagsOutput = "flags"
+	summaryOutput = "short"
+	yamlOutput    = "yaml"
+	jsonOutput    = "json"
+	flagsOutput   = "flags"
 )
--- a/pkg/cmd/hgctl/config_cluster.go
+++ b/pkg/cmd/hgctl/config_cluster.go
@@ -19,6 +19,7 @@ import (

 	"github.com/alibaba/higress/cmd/hgctl/config"
 	"github.com/spf13/cobra"
+	"istio.io/istio/istioctl/pkg/writer/envoy/configdump"
 	cmdutil "k8s.io/kubectl/pkg/cmd/util"
 )

@@ -49,17 +50,23 @@ func runClusterConfig(c *cobra.Command, args []string) error {
 	if len(args) != 0 {
 		podName = args[0]
 	}
-	envoyConfig, err := config.GetEnvoyConfig(&config.GetEnvoyConfigOptions{
+	configWriter, err := config.GetEnvoyConfigWriter(&config.GetEnvoyConfigOptions{
 		PodName:         podName,
 		PodNamespace:    podNamespace,
 		BindAddress:     bindAddress,
 		Output:          output,
 		EnvoyConfigType: config.ClusterEnvoyConfigType,
 		IncludeEds:      true,
-	})
+	}, c.OutOrStdout())
 	if err != nil {
 		return err
 	}
-	_, err = fmt.Fprintln(c.OutOrStdout(), string(envoyConfig))
-	return err
+	switch output {
+	case summaryOutput:
+		return configWriter.PrintClusterSummary(configdump.ClusterFilter{})
+	case jsonOutput, yamlOutput:
+		return configWriter.PrintClusterDump(configdump.ClusterFilter{}, output)
+	default:
+		return fmt.Errorf("output format %q not supported", output)
+	}
 }
--- a/pkg/cmd/hgctl/config_cmd.go
+++ b/pkg/cmd/hgctl/config_cmd.go
@@ -52,7 +52,7 @@ func newConfigCommand() *cobra.Command {
 	flags := cfgCommand.Flags()
 	options.AddKubeConfigFlags(flags)

-	cfgCommand.PersistentFlags().StringVarP(&output, "output", "o", "json", "One of 'yaml' or 'json'")
+	cfgCommand.PersistentFlags().StringVarP(&output, "output", "o", "json", "Output format: one of json|yaml|short")
 	cfgCommand.PersistentFlags().StringVarP(&podNamespace, "namespace", "n", "higress-system", "Namespace where envoy proxy pod are installed.")

 	return cfgCommand
--- a/pkg/cmd/hgctl/config_listener.go
+++ b/pkg/cmd/hgctl/config_listener.go
@@ -19,6 +19,7 @@ import (

 	"github.com/alibaba/higress/cmd/hgctl/config"
 	"github.com/spf13/cobra"
+	"istio.io/istio/istioctl/pkg/writer/envoy/configdump"
 	cmdutil "k8s.io/kubectl/pkg/cmd/util"
 )

@@ -49,17 +50,23 @@ func runListenerConfig(c *cobra.Command, args []string) error {
 	if len(args) != 0 {
 		podName = args[0]
 	}
-	envoyConfig, err := config.GetEnvoyConfig(&config.GetEnvoyConfigOptions{
+	configWriter, err := config.GetEnvoyConfigWriter(&config.GetEnvoyConfigOptions{
 		PodName:         podName,
 		PodNamespace:    podNamespace,
 		BindAddress:     bindAddress,
 		Output:          output,
 		EnvoyConfigType: config.ListenerEnvoyConfigType,
 		IncludeEds:      true,
-	})
+	}, c.OutOrStdout())
 	if err != nil {
 		return err
 	}
-	_, err = fmt.Fprintln(c.OutOrStdout(), string(envoyConfig))
-	return err
+	switch output {
+	case summaryOutput:
+		return configWriter.PrintListenerSummary(configdump.ListenerFilter{Verbose: true})
+	case jsonOutput, yamlOutput:
+		return configWriter.PrintListenerDump(configdump.ListenerFilter{Verbose: true}, output)
+	default:
+		return fmt.Errorf("output format %q not supported", output)
+	}
 }
--- a/pkg/cmd/hgctl/config_route.go
+++ b/pkg/cmd/hgctl/config_route.go
@@ -19,6 +19,7 @@ import (

 	"github.com/alibaba/higress/cmd/hgctl/config"
 	"github.com/spf13/cobra"
+	"istio.io/istio/istioctl/pkg/writer/envoy/configdump"
 	cmdutil "k8s.io/kubectl/pkg/cmd/util"
 )

@@ -49,17 +50,23 @@ func runRouteConfig(c *cobra.Command, args []string) error {
 	if len(args) != 0 {
 		podName = args[0]
 	}
-	envoyConfig, err := config.GetEnvoyConfig(&config.GetEnvoyConfigOptions{
+	configWriter, err := config.GetEnvoyConfigWriter(&config.GetEnvoyConfigOptions{
 		PodName:         podName,
 		PodNamespace:    podNamespace,
 		BindAddress:     bindAddress,
 		Output:          output,
 		EnvoyConfigType: config.RouteEnvoyConfigType,
 		IncludeEds:      true,
-	})
+	}, c.OutOrStdout())
 	if err != nil {
 		return err
 	}
-	_, err = fmt.Fprintln(c.OutOrStdout(), string(envoyConfig))
-	return err
+	switch output {
+	case summaryOutput:
+		return configWriter.PrintRouteSummary(configdump.RouteFilter{Verbose: true})
+	case jsonOutput, yamlOutput:
+		return configWriter.PrintRouteDump(configdump.RouteFilter{Verbose: true}, output)
+	default:
+		return fmt.Errorf("output format %q not supported", output)
+	}
 }
--- a/pkg/cmd/hgctl/dashboard.go
+++ b/pkg/cmd/hgctl/dashboard.go
@@ -425,7 +425,7 @@ func openCommand(writer io.Writer, command string, args ...string) {
 	_, err := exec.LookPath(command)
 	if err != nil {
 		if errors.Is(err, exec.ErrNotFound) {
-			fmt.Fprintf(writer, "Could not open your browser. Please open it maually.\n")
+			fmt.Fprintf(writer, "Could not open your browser. Please open it manually.\n")
 			return
 		}
 		fmt.Fprintf(writer, "Failed to open browser; open %s in your browser.\nError: %s\n", args[0], err.Error())
--- a/pkg/cmd/hgctl/install.go
+++ b/pkg/cmd/hgctl/install.go
@@ -28,7 +28,7 @@ import (

 const (
 	setFlagHelpStr = `Override an higress profile value, e.g. to choose a profile
-(--set profile=local-k8s), or override profile values (--set gateway.replicas=2), or override helm values (--set values.global.proxy.resources.requsts.cpu=500m).`
+(--set profile=local-k8s), or override profile values (--set gateway.replicas=2), or override helm values (--set values.global.proxy.resources.requests.cpu=500m).`
 	// manifestsFlagHelpStr is the command line description for --manifests
 	manifestsFlagHelpStr = `Specify a path to a directory of profiles
 (e.g. ~/Downloads/higress/manifests).`
@@ -101,7 +101,7 @@ func newInstallCmd() *cobra.Command {
  hgctl install --set profile=local-k8s  --set global.enableIstioAPI=true --set gateway.replicas=2"

  # To override helm setting
-  hgctl install --set profile=local-k8s  --set values.global.proxy.resources.requsts.cpu=500m"
+  hgctl install --set profile=local-k8s  --set values.global.proxy.resources.requests.cpu=500m"


 `,
@@ -175,7 +175,7 @@ func promptInstall(writer io.Writer, profileName string) bool {

 func promptProfileName(writer io.Writer) string {
 	answer := ""
-	fmt.Fprintf(writer, "\nPlease select higress install configration profile:\n")
+	fmt.Fprintf(writer, "\nPlease select higress install configuration profile:\n")
 	fmt.Fprintf(writer, "\n1.Install higress to local kubernetes cluster like kind etc.\n")
 	fmt.Fprintf(writer, "\n2.Install higress to kubernetes cluster\n")
 	fmt.Fprintf(writer, "\n3.Install higress to local docker environment\n")
--- a/pkg/cmd/hgctl/installer/standalone_agent.go
+++ b/pkg/cmd/hgctl/installer/standalone_agent.go
@@ -176,7 +176,7 @@ func (a *Agent) checkSudoPermission() error {
 	case <-time.After(5 * time.Second):
 		cmd2.Process.Signal(os.Interrupt)
 		if !a.quiet {
-			fmt.Fprintf(a.writer, "checked result: timeout execeed and need sudo with password\n")
+			fmt.Fprintf(a.writer, "checked result: timeout exceed and need sudo with password\n")
 		}
 		a.runSudoState = SudoWithPassword

--- a/pkg/cmd/hgctl/upgrade.go
+++ b/pkg/cmd/hgctl/upgrade.go
@@ -108,7 +108,7 @@ func upgrade(writer io.Writer, iArgs *InstallArgs) error {
 func promptUpgrade(writer io.Writer) bool {
 	answer := ""
 	for {
-		fmt.Fprintf(writer, "All Higress resources will be upgraed from the cluster. \nProceed? (y/N)")
+		fmt.Fprintf(writer, "All Higress resources will be upgrade from the cluster. \nProceed? (y/N)")
 		fmt.Scanln(&answer)
 		if strings.TrimSpace(answer) == "y" {
 			fmt.Fprintf(writer, "\n")
@@ -170,7 +170,7 @@ func promptProfileContexts(writer io.Writer, profileContexts []*installer.Profil
 	if len(profileContexts) == 1 {
 		fmt.Fprintf(writer, "\nFound a profile::  ")
 	} else {
-		fmt.Fprintf(writer, "\nPlease select higress installed configration profiles:\n")
+		fmt.Fprintf(writer, "\nPlease select higress installed configuration profiles:\n")
 	}
 	index := 1
 	for _, profileContext := range profileContexts {
--- a/pkg/common/protocol.go
+++ b/pkg/common/protocol.go
@@ -32,7 +32,7 @@ func ParseProtocol(s string) Protocol {
 		return TCP
 	case "http":
 		return HTTP
-	case "grpc":
+	case "grpc", "triple", "tri":
 		return GRPC
 	case "dubbo":
 		return Dubbo
--- a/pkg/ingress/config/ingress_config.go
+++ b/pkg/ingress/config/ingress_config.go
@@ -841,6 +841,7 @@ func (m *IngressConfig) convertIstioWasmPlugin(obj *higressext.WasmPlugin) (*ext
 				StructValue: rule.Config,
 			}
 			var matchItems []*types.Value
+			// match ingress
 			for _, ing := range rule.Ingress {
 				matchItems = append(matchItems, &types.Value{
 					Kind: &types.Value_StringValue{
@@ -861,6 +862,7 @@ func (m *IngressConfig) convertIstioWasmPlugin(obj *higressext.WasmPlugin) (*ext
 				})
 				continue
 			}
+			// match domain
 			for _, domain := range rule.Domain {
 				matchItems = append(matchItems, &types.Value{
 					Kind: &types.Value_StringValue{
@@ -868,10 +870,31 @@ func (m *IngressConfig) convertIstioWasmPlugin(obj *higressext.WasmPlugin) (*ext
 					},
 				})
 			}
+			if len(matchItems) > 0 {
+				v.StructValue.Fields["_match_domain_"] = &types.Value{
+					Kind: &types.Value_ListValue{
+						ListValue: &types.ListValue{
+							Values: matchItems,
+						},
+					},
+				}
+				ruleValues = append(ruleValues, &types.Value{
+					Kind: v,
+				})
+				continue
+			}
+			// match service
+			for _, service := range rule.Service {
+				matchItems = append(matchItems, &types.Value{
+					Kind: &types.Value_StringValue{
+						StringValue: service,
+					},
+				})
+			}
 			if len(matchItems) == 0 {
 				return nil, fmt.Errorf("invalid match rule has no match condition, rule:%v", rule)
 			}
-			v.StructValue.Fields["_match_domain_"] = &types.Value{
+			v.StructValue.Fields["_match_service_"] = &types.Value{
 				Kind: &types.Value_ListValue{
 					ListValue: &types.ListValue{
 						Values: matchItems,
@@ -918,7 +941,7 @@ func (m *IngressConfig) AddOrUpdateWasmPlugin(clusterNamespacedName util.Cluster
 		Labels: map[string]string{constants.AlwaysPushLabel: "true"},
 	}
 	for _, f := range m.wasmPluginHandlers {
-		IngressLog.Debug("WasmPlugin triggerd update")
+		IngressLog.Debug("WasmPlugin triggered update")
 		f(config.Config{Meta: metadata}, config.Config{Meta: metadata}, model.EventUpdate)
 	}
 	istioWasmPlugin, err := m.convertIstioWasmPlugin(&wasmPlugin.Spec)
@@ -960,7 +983,7 @@ func (m *IngressConfig) DeleteWasmPlugin(clusterNamespacedName util.ClusterNames
 			Labels: map[string]string{constants.AlwaysPushLabel: "true"},
 		}
 		for _, f := range m.wasmPluginHandlers {
-			IngressLog.Debug("WasmPlugin triggerd update")
+			IngressLog.Debug("WasmPlugin triggered update")
 			f(config.Config{Meta: metadata}, config.Config{Meta: metadata}, model.EventDelete)
 		}
 	}
@@ -987,7 +1010,7 @@ func (m *IngressConfig) AddOrUpdateMcpBridge(clusterNamespacedName util.ClusterN
 				Labels: map[string]string{constants.AlwaysPushLabel: "true"},
 			}
 			for _, f := range m.serviceEntryHandlers {
-				IngressLog.Debug("McpBridge triggerd serviceEntry update")
+				IngressLog.Debug("McpBridge triggered serviceEntry update")
 				f(config.Config{Meta: metadata}, config.Config{Meta: metadata}, model.EventUpdate)
 			}
 		}, m.localKubeClient, m.namespace)
@@ -1042,7 +1065,7 @@ func (m *IngressConfig) AddOrUpdateHttp2Rpc(clusterNamespacedName util.ClusterNa
 }

 func (m *IngressConfig) DeleteHttp2Rpc(clusterNamespacedName util.ClusterNamespacedName) {
-	IngressLog.Infof("Http2Rpc triggerd deleted event %s", clusterNamespacedName.Name)
+	IngressLog.Infof("Http2Rpc triggered deleted event %s", clusterNamespacedName.Name)
 	if clusterNamespacedName.Namespace != m.namespace {
 		return
 	}
@@ -1054,7 +1077,7 @@ func (m *IngressConfig) DeleteHttp2Rpc(clusterNamespacedName util.ClusterNamespa
 	}
 	m.mutex.Unlock()
 	if hit {
-		IngressLog.Infof("Http2Rpc triggerd deleted event executed %s", clusterNamespacedName.Name)
+		IngressLog.Infof("Http2Rpc triggered deleted event executed %s", clusterNamespacedName.Name)
 		push := func(kind config.GroupVersionKind) {
 			m.XDSUpdater.ConfigUpdate(&model.PushRequest{
 				Full: true,
@@ -1160,13 +1183,13 @@ func (m *IngressConfig) constructHttp2RpcEnvoyFilter(http2rpcConfig *annotations
 	IngressLog.Infof("Found http2rpc mappings %v", mappings)
 	if _, exist := mappings[http2rpcConfig.Name]; !exist {
 		IngressLog.Errorf("Http2RpcConfig name %s, not found Http2Rpc CRD", http2rpcConfig.Name)
-		return nil, errors.New("invalid http2rpcConfig has no useable http2rpc")
+		return nil, errors.New("invalid http2rpcConfig has no usable http2rpc")
 	}
 	http2rpcCRD := mappings[http2rpcConfig.Name]

 	if http2rpcCRD.GetDubbo() == nil {
 		IngressLog.Errorf("Http2RpcConfig name %s, only support Http2Rpc CRD Dubbo Service type", http2rpcConfig.Name)
-		return nil, errors.New("invalid http2rpcConfig has no useable http2rpc")
+		return nil, errors.New("invalid http2rpcConfig has no usable http2rpc")
 	}

 	httpRoute := route.HTTPRoute
@@ -1293,7 +1316,7 @@ func (m *IngressConfig) constructHttp2RpcMethods(dubbo *higressv1.DubboService)
 		var method = make(map[string]interface{})
 		method["name"] = serviceMethod.GetServiceMethod()
 		var params []interface{}
-		// paramFromEntireBody is for methods with single parameter. So when paramFromEntireBody exists, we just ignore parmas.
+		// paramFromEntireBody is for methods with single parameter. So when paramFromEntireBody exists, we just ignore params.
 		var paramFromEntireBody = serviceMethod.GetParamFromEntireBody()
 		if paramFromEntireBody != nil {
 			var param = make(map[string]interface{})
--- a/pkg/ingress/kube/ingress/controller.go
+++ b/pkg/ingress/kube/ingress/controller.go
@@ -433,6 +433,11 @@ func (c *controller) ConvertGateway(convertOptions *common.ConvertOptions, wrapp
 						// If there is no matching secret, try to get it from configmap.
 						secretName = httpsCredentialConfig.MatchSecretNameByDomain(rule.Host)
 						secretNamespace = c.options.SystemNamespace
+						namespace, secret := cert.ParseTLSSecret(secretName)
+						if namespace != "" {
+							secretNamespace = namespace
+							secretName = secret
+						}
 					}
 				}
 			}
@@ -441,6 +446,11 @@ func (c *controller) ConvertGateway(convertOptions *common.ConvertOptions, wrapp
 			if httpsCredentialConfig != nil {
 				secretName = httpsCredentialConfig.MatchSecretNameByDomain(rule.Host)
 				secretNamespace = c.options.SystemNamespace
+				namespace, secret := cert.ParseTLSSecret(secretName)
+				if namespace != "" {
+					secretNamespace = namespace
+					secretName = secret
+				}
 			}
 		}
 		if secretName == "" {
--- a/pkg/ingress/kube/ingressv1/controller.go
+++ b/pkg/ingress/kube/ingressv1/controller.go
@@ -419,6 +419,11 @@ func (c *controller) ConvertGateway(convertOptions *common.ConvertOptions, wrapp
 						// If there is no matching secret, try to get it from configmap.
 						secretName = httpsCredentialConfig.MatchSecretNameByDomain(rule.Host)
 						secretNamespace = c.options.SystemNamespace
+						namespace, secret := cert.ParseTLSSecret(secretName)
+						if namespace != "" {
+							secretNamespace = namespace
+							secretName = secret
+						}
 					}
 				}
 			}
@@ -427,6 +432,11 @@ func (c *controller) ConvertGateway(convertOptions *common.ConvertOptions, wrapp
 			if httpsCredentialConfig != nil {
 				secretName = httpsCredentialConfig.MatchSecretNameByDomain(rule.Host)
 				secretNamespace = c.options.SystemNamespace
+				namespace, secret := cert.ParseTLSSecret(secretName)
+				if namespace != "" {
+					secretNamespace = namespace
+					secretName = secret
+				}
 			}
 		}

--- a/plugins/wasm-go/Makefile
+++ b/plugins/wasm-go/Makefile
@@ -5,7 +5,7 @@ GO_VERSION ?= 1.19
 TINYGO_VERSION ?= 0.28.1
 ORAS_VERSION ?= 1.0.0
 HIGRESS_VERSION ?= 1.0.0-rc
-USE_HIGRESS_TINYGO ?= true
+USE_HIGRESS_TINYGO ?= false
 BUILDER ?= ${BUILDER_REGISTRY}wasm-go-builder:go${GO_VERSION}-tinygo${TINYGO_VERSION}-oras${ORAS_VERSION}
 BUILD_TIME := $(shell date "+%Y%m%d-%H%M%S")
 COMMIT_ID := $(shell git rev-parse --short HEAD 2>/dev/null)
--- a/plugins/wasm-go/extensions/ai-cache/.buildrc
+++ b/plugins/wasm-go/extensions/ai-cache/.buildrc
@@ -0,0 +1 @@
+EXTRA_TAGS=proxy_wasm_version_0_2_100
--- a/plugins/wasm-go/extensions/ai-cache/README.md
+++ b/plugins/wasm-go/extensions/ai-cache/README.md
@@ -32,3 +32,15 @@ redis:
  serviceName: my-redis.dns
  timeout: 2000
 ```
+
+## 进阶用法
+
+当前默认的缓存 key 是基于 GJSON PATH 的表达式：`messages.@reverse.0.content` 提取，含义是把 messages 数组反转后取第一项的 content；
+
+GJSON PATH 支持条件判断语法，例如希望取最后一个 role 为 user 的 content 作为 key，可以写成： `messages.@reverse.#(role=="user").content`；
+
+如果希望将所有 role 为 user 的 content 拼成一个数组作为 key，可以写成：`messages.@reverse.#(role=="user")#.content`；
+
+还可以支持管道语法，例如希望取到数第二个 role 为 user 的 content 作为 key，可以写成：`messages.@reverse.#(role=="user")#.content|1`。
+
+更多用法可以参考[官方文档](https://github.com/tidwall/gjson/blob/master/SYNTAX.md)，可以使用 [GJSON Playground](https://gjson.dev/) 进行语法测试。
--- a/plugins/wasm-go/extensions/ai-cache/go.mod
+++ b/plugins/wasm-go/extensions/ai-cache/go.mod
@@ -8,7 +8,7 @@ replace github.com/alibaba/higress/plugins/wasm-go => ../..

 require (
 	github.com/alibaba/higress/plugins/wasm-go v1.3.6-0.20240528060522-53bccf89f441
-	github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240327114451-d6b7174a84fc
+	github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240711023527-ba358c48772f
 	github.com/tidwall/gjson v1.14.3
 	github.com/tidwall/resp v0.1.1
 	github.com/tidwall/sjson v1.2.5
--- a/plugins/wasm-go/extensions/ai-cache/go.sum
+++ b/plugins/wasm-go/extensions/ai-cache/go.sum
@@ -5,6 +5,7 @@ github.com/higress-group/nottinygc v0.0.0-20231101025119-e93c4c2f8520 h1:IHDghbG
 github.com/higress-group/nottinygc v0.0.0-20231101025119-e93c4c2f8520/go.mod h1:Nz8ORLaFiLWotg6GeKlJMhv8cci8mM43uEnLA5t8iew=
 github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240327114451-d6b7174a84fc h1:t2AT8zb6N/59Y78lyRWedVoVWHNRSCBh0oWCC+bluTQ=
 github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240327114451-d6b7174a84fc/go.mod h1:hNFjhrLUIq+kJ9bOcs8QtiplSQ61GZXtd2xHKx4BYRo=
+github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240711023527-ba358c48772f/go.mod h1:hNFjhrLUIq+kJ9bOcs8QtiplSQ61GZXtd2xHKx4BYRo=
 github.com/magefile/mage v1.14.0 h1:6QDX3g6z1YvJ4olPhT1wksUcSa/V0a1B+pJb73fBjyo=
 github.com/magefile/mage v1.14.0/go.mod h1:z5UZb/iS3GoOSn0JgWuiw7dxlurVYTu+/jHXqQg881A=
 github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
--- a/plugins/wasm-go/extensions/ai-cache/main.go
+++ b/plugins/wasm-go/extensions/ai-cache/main.go
@@ -222,9 +222,9 @@ func onHttpRequestBody(ctx wrapper.HttpContext, config PluginConfig, body []byte
 		log.Debugf("cache hit, key:%s", key)
 		ctx.SetContext(CacheKeyContextKey, nil)
 		if !stream {
-			proxywasm.SendHttpResponse(200, [][2]string{{"content-type", "application/json; charset=utf-8"}}, []byte(fmt.Sprintf(config.ReturnResponseTemplate, response.String())), -1)
+			proxywasm.SendHttpResponseWithDetail(200, "ai-cache.hit", [][2]string{{"content-type", "application/json; charset=utf-8"}}, []byte(fmt.Sprintf(config.ReturnResponseTemplate, response.String())), -1)
 		} else {
-			proxywasm.SendHttpResponse(200, [][2]string{{"content-type", "text/event-stream; charset=utf-8"}}, []byte(fmt.Sprintf(config.ReturnStreamResponseTemplate, response.String())), -1)
+			proxywasm.SendHttpResponseWithDetail(200, "ai-cache.hit", [][2]string{{"content-type", "text/event-stream; charset=utf-8"}}, []byte(fmt.Sprintf(config.ReturnStreamResponseTemplate, response.String())), -1)
 		}
 	})
 	if err != nil {
--- a/plugins/wasm-go/extensions/ai-prompt-decorator/.gitignore
+++ b/plugins/wasm-go/extensions/ai-prompt-decorator/.gitignore
@@ -0,0 +1,3 @@
+config.yaml
+main.wasm
+tmp/
--- a/plugins/wasm-go/extensions/ai-prompt-decorator/README.md
+++ b/plugins/wasm-go/extensions/ai-prompt-decorator/README.md
@@ -0,0 +1,69 @@
+# 简介
+AI提示词装饰器插件，支持在LLM的请求前后插入prompt。
+
+# 配置说明
+
+| 名称 | 数据类型 | 填写要求 | 默认值 | 描述 |
+|----------------|-----------------|------|-----|----------------------------------|
+| `prepend` | array of message object | optional | - | 在初始输入之前插入的语句 |
+| `append` | array of message object | optional | - | 在初始输入之后插入的语句 |
+
+message object 配置说明：
+
+| 名称 | 数据类型 | 填写要求 | 默认值 | 描述 |
+|----------------|-----------------|------|-----|----------------------------------|
+| `role` | string | 必填 | - | 角色 |
+| `content` | string | 必填 | - | 消息 |
+
+# 示例
+
+配置示例如下：
+
+```yaml
+prepend:
+- role: system
+  content: "请使用英语回答问题"
+append:
+- role: user
+  content: "每次回答完问题，尝试进行反问"
+```
+
+使用以上配置发起请求：
+
+```bash
+curl http://localhost/test \
+-H "content-type: application/json" \
+-d '{
+  "model": "gpt-3.5-turbo",
+  "messages": [
+    {
+      "role": "user",
+      "content": "你是谁？"
+    }
+  ]
+}
+```
+
+经过插件处理后，实际请求为：
+
+```bash
+curl http://localhost/test \
+-H "content-type: application/json" \
+-d '{
+  "model": "gpt-3.5-turbo",
+  "messages": [
+    {
+      "role": "system",
+      "content": "请使用英语回答问题"
+    },
+    {
+      "role": "user",
+      "content": "你是谁？"
+    },
+    {
+      "role": "user",
+      "content": "每次回答完问题，尝试进行反问"
+    }
+  ]
+}
+```
--- a/plugins/wasm-go/extensions/ai-prompt-decorator/go.mod
+++ b/plugins/wasm-go/extensions/ai-prompt-decorator/go.mod
@@ -0,0 +1,21 @@
+module ai-prompt-decorator
+
+go 1.18
+
+replace github.com/alibaba/higress/plugins/wasm-go => ../..
+
+require (
+	github.com/alibaba/higress/plugins/wasm-go v1.3.5
+	github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240711023527-ba358c48772f
+	github.com/tidwall/gjson v1.14.3
+)
+
+require (
+	github.com/google/uuid v1.3.0 // indirect
+	github.com/higress-group/nottinygc v0.0.0-20231101025119-e93c4c2f8520 // indirect
+	github.com/magefile/mage v1.14.0 // indirect
+	github.com/tidwall/match v1.1.1 // indirect
+	github.com/tidwall/pretty v1.2.0 // indirect
+	github.com/tidwall/resp v0.1.1 // indirect
+	github.com/tidwall/sjson v1.2.5
+)
--- a/plugins/wasm-go/extensions/ai-prompt-decorator/go.sum
+++ b/plugins/wasm-go/extensions/ai-prompt-decorator/go.sum
@@ -0,0 +1,23 @@
+github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
+github.com/google/uuid v1.3.0 h1:t6JiXgmwXMjEs8VusXIJk2BXHsn+wx8BZdTaoZ5fu7I=
+github.com/google/uuid v1.3.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
+github.com/higress-group/nottinygc v0.0.0-20231101025119-e93c4c2f8520 h1:IHDghbGQ2DTIXHBHxWfqCYQW1fKjyJ/I7W1pMyUDeEA=
+github.com/higress-group/nottinygc v0.0.0-20231101025119-e93c4c2f8520/go.mod h1:Nz8ORLaFiLWotg6GeKlJMhv8cci8mM43uEnLA5t8iew=
+github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240711023527-ba358c48772f h1:ZIiIBRvIw62gA5MJhuwp1+2wWbqL9IGElQ499rUsYYg=
+github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240711023527-ba358c48772f/go.mod h1:hNFjhrLUIq+kJ9bOcs8QtiplSQ61GZXtd2xHKx4BYRo=
+github.com/magefile/mage v1.14.0 h1:6QDX3g6z1YvJ4olPhT1wksUcSa/V0a1B+pJb73fBjyo=
+github.com/magefile/mage v1.14.0/go.mod h1:z5UZb/iS3GoOSn0JgWuiw7dxlurVYTu+/jHXqQg881A=
+github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
+github.com/stretchr/testify v1.8.4 h1:CcVxjf3Q8PM0mHUKJCdn+eZZtm5yQwehR5yeSVQQcUk=
+github.com/tidwall/gjson v1.14.2/go.mod h1:/wbyibRr2FHMks5tjHJ5F8dMZh3AcwJEMf5vlfC0lxk=
+github.com/tidwall/gjson v1.14.3 h1:9jvXn7olKEHU1S9vwoMGliaT8jq1vJ7IH/n9zD9Dnlw=
+github.com/tidwall/gjson v1.14.3/go.mod h1:/wbyibRr2FHMks5tjHJ5F8dMZh3AcwJEMf5vlfC0lxk=
+github.com/tidwall/match v1.1.1 h1:+Ho715JplO36QYgwN9PGYNhgZvoUSc9X2c80KVTi+GA=
+github.com/tidwall/match v1.1.1/go.mod h1:eRSPERbgtNPcGhD8UCthc6PmLEQXEWd3PRB5JTxsfmM=
+github.com/tidwall/pretty v1.2.0 h1:RWIZEg2iJ8/g6fDDYzMpobmaoGh5OLl4AXtGUGPcqCs=
+github.com/tidwall/pretty v1.2.0/go.mod h1:ITEVvHYasfjBbM0u2Pg8T2nJnzm8xPwvNhhsoaGGjNU=
+github.com/tidwall/resp v0.1.1 h1:Ly20wkhqKTmDUPlyM1S7pWo5kk0tDu8OoC/vFArXmwE=
+github.com/tidwall/resp v0.1.1/go.mod h1:3/FrruOBAxPTPtundW0VXgmsQ4ZBA0Aw714lVYgwFa0=
+github.com/tidwall/sjson v1.2.5 h1:kLy8mja+1c9jlljvWTlSazM7cKDRfJuR/bOJhcY5NcY=
+github.com/tidwall/sjson v1.2.5/go.mod h1:Fvgq9kS/6ociJEDnK0Fk1cpYF4FIW6ZF7LAe+6jwd28=
+gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
--- a/plugins/wasm-go/extensions/ai-prompt-decorator/main.go
+++ b/plugins/wasm-go/extensions/ai-prompt-decorator/main.go
@@ -0,0 +1,80 @@
+package main
+
+import (
+	"encoding/json"
+
+	"github.com/alibaba/higress/plugins/wasm-go/pkg/wrapper"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm/types"
+	"github.com/tidwall/gjson"
+	"github.com/tidwall/sjson"
+)
+
+func main() {
+	wrapper.SetCtx(
+		"ai-prompt-decorator",
+		wrapper.ParseConfigBy(parseConfig),
+		wrapper.ProcessRequestHeadersBy(onHttpRequestHeaders),
+		wrapper.ProcessRequestBodyBy(onHttpRequestBody),
+	)
+}
+
+type Message struct {
+	Role    string `json:"role"`
+	Content string `json:"content"`
+}
+
+type AIPromptDecoratorConfig struct {
+	Prepend []Message `json:"prepend"`
+	Append  []Message `json:"append"`
+}
+
+func parseConfig(jsonConfig gjson.Result, config *AIPromptDecoratorConfig, log wrapper.Log) error {
+	return json.Unmarshal([]byte(jsonConfig.Raw), config)
+}
+
+func onHttpRequestHeaders(ctx wrapper.HttpContext, config AIPromptDecoratorConfig, log wrapper.Log) types.Action {
+	proxywasm.RemoveHttpRequestHeader("content-length")
+	return types.ActionContinue
+}
+
+func onHttpRequestBody(ctx wrapper.HttpContext, config AIPromptDecoratorConfig, body []byte, log wrapper.Log) types.Action {
+	messageJson := `{"messages":[]}`
+
+	for _, entry := range config.Prepend {
+		msg, err := json.Marshal(entry)
+		if err != nil {
+			log.Errorf("Failed to add prepend message, error: %v", err)
+			return types.ActionContinue
+		}
+		messageJson, _ = sjson.SetRaw(messageJson, "messages.-1", string(msg))
+	}
+
+	rawMessage := gjson.GetBytes(body, "messages")
+	if !rawMessage.Exists() {
+		log.Errorf("Cannot find messages field in request body")
+		return types.ActionContinue
+	}
+	for _, entry := range rawMessage.Array() {
+		messageJson, _ = sjson.SetRaw(messageJson, "messages.-1", entry.Raw)
+	}
+
+	for _, entry := range config.Append {
+		msg, err := json.Marshal(entry)
+		if err != nil {
+			log.Errorf("Failed to add prepend message, error: %v", err)
+			return types.ActionContinue
+		}
+		messageJson, _ = sjson.SetRaw(messageJson, "messages.-1", string(msg))
+	}
+
+	newbody, err := sjson.SetRaw(string(body), "messages", gjson.Get(messageJson, "messages").Raw)
+	if err != nil {
+		log.Error("modify body failed")
+	}
+	if err = proxywasm.ReplaceHttpRequestBody([]byte(newbody)); err != nil {
+		log.Error("rewrite body failed")
+	}
+
+	return types.ActionContinue
+}
--- a/plugins/wasm-go/extensions/ai-prompt-template/.gitignore
+++ b/plugins/wasm-go/extensions/ai-prompt-template/.gitignore
@@ -0,0 +1,3 @@
+config.yaml
+main.wasm
+tmp/
--- a/plugins/wasm-go/extensions/ai-prompt-template/README.md
+++ b/plugins/wasm-go/extensions/ai-prompt-template/README.md
@@ -0,0 +1,48 @@
+# 简介
+AI提示词模板，用于快速构建同类型的AI请求。
+
+# 配置说明
+| 名称 | 数据类型 | 填写要求 | 默认值 | 描述 |
+|----------------|-----------------|------|-----|----------------------------------|
+| `templates` | array of object | 必填 | - | 模板设置 |
+
+template object 配置说明：
+
+| 名称 | 数据类型 | 填写要求 | 默认值 | 描述 |
+|----------------|-----------------|------|-----|----------------------------------|
+| `name` | string | 必填 | - | 模板名称 |
+| `template.model` | string | 必填 | - | 模型名称 |
+| `template.messages` | array of object | 必填 | - | 大模型输入 |
+
+message object 配置说明：
+
+| 名称 | 数据类型 | 填写要求 | 默认值 | 描述 |
+|----------------|-----------------|------|-----|----------------------------------|
+| `role` | string | 必填 | - | 角色 |
+| `content` | string | 必填 | - | 消息 |
+
+配置示例如下：
+
+```yaml
+templates:
+- name: "developer-chat"
+  template:
+    model: gpt-3.5-turbo
+    messages:
+    - role: system
+      content: "You are a {{program}} expert, in {{language}} programming language."
+    - role: user
+      content: "Write me a {{program}} program."
+```
+
+使用以上配置的请求body示例：
+
+```json
+{
+  "template": "developer-chat",
+  "properties": {
+    "program": "quick sort",
+    "language": "python"
+  }
+}
+```
--- a/plugins/wasm-go/extensions/ai-prompt-template/go.mod
+++ b/plugins/wasm-go/extensions/ai-prompt-template/go.mod
@@ -0,0 +1,21 @@
+module ai-prompt-template
+
+go 1.18
+
+replace github.com/alibaba/higress/plugins/wasm-go => ../..
+
+require (
+	github.com/alibaba/higress/plugins/wasm-go v1.3.5
+	github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240711023527-ba358c48772f
+	github.com/tidwall/gjson v1.14.3
+)
+
+require (
+	github.com/google/uuid v1.3.0 // indirect
+	github.com/higress-group/nottinygc v0.0.0-20231101025119-e93c4c2f8520 // indirect
+	github.com/magefile/mage v1.14.0 // indirect
+	github.com/tidwall/match v1.1.1 // indirect
+	github.com/tidwall/pretty v1.2.0 // indirect
+	github.com/tidwall/resp v0.1.1 // indirect
+	github.com/tidwall/sjson v1.2.5
+)
--- a/plugins/wasm-go/extensions/ai-prompt-template/go.sum
+++ b/plugins/wasm-go/extensions/ai-prompt-template/go.sum
@@ -0,0 +1,26 @@
+github.com/alibaba/higress/plugins/wasm-go v1.3.5 h1:VOLL3m442IHCSu8mR5AZ4sc6LVT9X0w1hdqDI7oB9jY=
+github.com/alibaba/higress/plugins/wasm-go v1.3.5/go.mod h1:kr3V9Ntbspj1eSrX8rgjBsdMXkGupYEf+LM72caGPQc=
+github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
+github.com/google/uuid v1.3.0 h1:t6JiXgmwXMjEs8VusXIJk2BXHsn+wx8BZdTaoZ5fu7I=
+github.com/google/uuid v1.3.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
+github.com/higress-group/nottinygc v0.0.0-20231101025119-e93c4c2f8520 h1:IHDghbGQ2DTIXHBHxWfqCYQW1fKjyJ/I7W1pMyUDeEA=
+github.com/higress-group/nottinygc v0.0.0-20231101025119-e93c4c2f8520/go.mod h1:Nz8ORLaFiLWotg6GeKlJMhv8cci8mM43uEnLA5t8iew=
+github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240226064518-b3dc4646a35a h1:luYRvxLTE1xYxrXYj7nmjd1U0HHh8pUPiKfdZ0MhCGE=
+github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240226064518-b3dc4646a35a/go.mod h1:hNFjhrLUIq+kJ9bOcs8QtiplSQ61GZXtd2xHKx4BYRo=
+github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240711023527-ba358c48772f/go.mod h1:hNFjhrLUIq+kJ9bOcs8QtiplSQ61GZXtd2xHKx4BYRo=
+github.com/magefile/mage v1.14.0 h1:6QDX3g6z1YvJ4olPhT1wksUcSa/V0a1B+pJb73fBjyo=
+github.com/magefile/mage v1.14.0/go.mod h1:z5UZb/iS3GoOSn0JgWuiw7dxlurVYTu+/jHXqQg881A=
+github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
+github.com/stretchr/testify v1.8.4 h1:CcVxjf3Q8PM0mHUKJCdn+eZZtm5yQwehR5yeSVQQcUk=
+github.com/tidwall/gjson v1.14.2/go.mod h1:/wbyibRr2FHMks5tjHJ5F8dMZh3AcwJEMf5vlfC0lxk=
+github.com/tidwall/gjson v1.14.3 h1:9jvXn7olKEHU1S9vwoMGliaT8jq1vJ7IH/n9zD9Dnlw=
+github.com/tidwall/gjson v1.14.3/go.mod h1:/wbyibRr2FHMks5tjHJ5F8dMZh3AcwJEMf5vlfC0lxk=
+github.com/tidwall/match v1.1.1 h1:+Ho715JplO36QYgwN9PGYNhgZvoUSc9X2c80KVTi+GA=
+github.com/tidwall/match v1.1.1/go.mod h1:eRSPERbgtNPcGhD8UCthc6PmLEQXEWd3PRB5JTxsfmM=
+github.com/tidwall/pretty v1.2.0 h1:RWIZEg2iJ8/g6fDDYzMpobmaoGh5OLl4AXtGUGPcqCs=
+github.com/tidwall/pretty v1.2.0/go.mod h1:ITEVvHYasfjBbM0u2Pg8T2nJnzm8xPwvNhhsoaGGjNU=
+github.com/tidwall/resp v0.1.1 h1:Ly20wkhqKTmDUPlyM1S7pWo5kk0tDu8OoC/vFArXmwE=
+github.com/tidwall/resp v0.1.1/go.mod h1:3/FrruOBAxPTPtundW0VXgmsQ4ZBA0Aw714lVYgwFa0=
+github.com/tidwall/sjson v1.2.5 h1:kLy8mja+1c9jlljvWTlSazM7cKDRfJuR/bOJhcY5NcY=
+github.com/tidwall/sjson v1.2.5/go.mod h1:Fvgq9kS/6ociJEDnK0Fk1cpYF4FIW6ZF7LAe+6jwd28=
+gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
--- a/plugins/wasm-go/extensions/ai-prompt-template/main.go
+++ b/plugins/wasm-go/extensions/ai-prompt-template/main.go
@@ -0,0 +1,55 @@
+package main
+
+import (
+	"fmt"
+	"strings"
+
+	"github.com/alibaba/higress/plugins/wasm-go/pkg/wrapper"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm/types"
+	"github.com/tidwall/gjson"
+)
+
+func main() {
+	wrapper.SetCtx(
+		"ai-prompt-template",
+		wrapper.ParseConfigBy(parseConfig),
+		wrapper.ProcessRequestHeadersBy(onHttpRequestHeaders),
+		wrapper.ProcessRequestBodyBy(onHttpRequestBody),
+	)
+}
+
+type AIPromptTemplateConfig struct {
+	templates map[string]string
+}
+
+func parseConfig(json gjson.Result, config *AIPromptTemplateConfig, log wrapper.Log) error {
+	config.templates = make(map[string]string)
+	for _, v := range json.Get("templates").Array() {
+		config.templates[v.Get("name").String()] = v.Get("template").Raw
+		log.Info(v.Get("template").Raw)
+	}
+	return nil
+}
+
+func onHttpRequestHeaders(ctx wrapper.HttpContext, config AIPromptTemplateConfig, log wrapper.Log) types.Action {
+	templateEnable, _ := proxywasm.GetHttpRequestHeader("template-enable")
+	if templateEnable != "true" {
+		ctx.DontReadRequestBody()
+		return types.ActionContinue
+	}
+	proxywasm.RemoveHttpRequestHeader("content-length")
+	return types.ActionContinue
+}
+
+func onHttpRequestBody(ctx wrapper.HttpContext, config AIPromptTemplateConfig, body []byte, log wrapper.Log) types.Action {
+	if gjson.GetBytes(body, "template").Exists() && gjson.GetBytes(body, "properties").Exists() {
+		name := gjson.GetBytes(body, "template").String()
+		template := config.templates[name]
+		for key, value := range gjson.GetBytes(body, "properties").Map() {
+			template = strings.ReplaceAll(template, fmt.Sprintf("{{%s}}", key), value.String())
+		}
+		proxywasm.ReplaceHttpRequestBody([]byte(template))
+	}
+	return types.ActionContinue
+}
--- a/plugins/wasm-go/extensions/ai-proxy/.buildrc
+++ b/plugins/wasm-go/extensions/ai-proxy/.buildrc
@@ -0,0 +1 @@
+EXTRA_TAGS=proxy_wasm_version_0_2_100
--- a/plugins/wasm-go/extensions/ai-proxy/README.md
+++ b/plugins/wasm-go/extensions/ai-proxy/README.md
@@ -1,6 +1,6 @@
 ---
 title: AI 代理
-keywords: [ higress,ai,proxy,rag ]
+keywords: [ AI网关, AI代理 ]
 description: AI 代理插件配置参考
 ---

@@ -9,6 +9,13 @@ description: AI 代理插件配置参考
 `AI 代理`插件实现了基于 OpenAI API 契约的 AI 代理功能。目前支持 OpenAI、Azure OpenAI、月之暗面（Moonshot）和通义千问等 AI
 服务提供商。

+> **注意：**
+
+> 请求路径后缀匹配 `/v1/chat/completions` 时，对应文生文场景，会用 OpenAI 的文生文协议解析请求 Body，再转换为对应 LLM 厂商的文生文协议
+
+> 请求路径后缀匹配 `/v1/embeddings` 时，对应文本向量场景，会用 OpenAI 的文本向量协议解析请求 Body，再转换为对应 LLM 厂商的文本向量协议
+
+
 ## 配置字段

 ### 基本配置
@@ -19,14 +26,14 @@ description: AI 代理插件配置参考

 `provider`的配置字段说明如下：

-| 名称             | 数据类型            | 填写要求 | 默认值 | 描述                                                                               |
-|----------------|-----------------|------|-----|----------------------------------------------------------------------------------|
-| `type`         | string          | 必填   | -   | AI 服务提供商名称。目前支持以下取值：openai, azure, moonshot, qwen, zhipuai                                |
-| `apiTokens`    | array of string | 必填   | -   | 用于在访问 AI 服务时进行认证的令牌。如果配置了多个 token，插件会在请求时随机进行选择。部分服务提供商只支持配置一个 token。            |
-| `timeout`      | number          | 非必填  | -   | 访问 AI 服务的超时时间。单位为毫秒。默认值为 120000，即 2 分钟                                           |
-| `modelMapping` | map of string   | 非必填  | -   | AI 模型映射表，用于将请求中的模型名称映射为服务提供商支持模型名称。<br/>可以使用 "*" 为键来配置通用兜底映射关系                   |
-| `protocol`     | string          | 非必填  | -   | 插件对外提供的 API 接口契约。目前支持以下取值：openai（默认值，使用 OpenAI 的接口契约）、original（使用目标服务提供商的原始接口契约） |
-| `context`      | object          | 非必填  | -   | 配置 AI 对话上下文信息                                                                    |
+| 名称           | 数据类型        | 填写要求 | 默认值 | 描述                                                                                                                                                                                                                                                           |
+| -------------- | --------------- | -------- | ------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------                                                                                                  |
+| `type`         | string          | 必填     | -      | AI 服务提供商名称                                                                                                                                                                                                                                              |
+| `apiTokens`    | array of string | 非必填   | -      | 用于在访问 AI 服务时进行认证的令牌。如果配置了多个 token，插件会在请求时随机进行选择。部分服务提供商只支持配置一个 token。                                                                                                                                     |
+| `timeout`      | number          | 非必填   | -      | 访问 AI 服务的超时时间。单位为毫秒。默认值为 120000，即 2 分钟                                                                                                                                                                                                 |
+| `modelMapping` | map of string   | 非必填   | -      | AI 模型映射表，用于将请求中的模型名称映射为服务提供商支持模型名称。<br/>1. 支持前缀匹配。例如用 "gpt-3-*" 匹配所有名称以“gpt-3-”开头的模型；<br/>2. 支持使用 "*" 为键来配置通用兜底映射关系；<br/>3. 如果映射的目标名称为空字符串 ""，则表示保留原模型名称。 |
+| `protocol`     | string          | 非必填   | -      | 插件对外提供的 API 接口契约。目前支持以下取值：openai（默认值，使用 OpenAI 的接口契约）、original（使用目标服务提供商的原始接口契约）                                                                                                                          |
+| `context`      | object          | 非必填   | -      | 配置 AI 对话上下文信息                                                                                                                                                                                                                                         |

 `context`的配置字段说明如下：

@@ -40,7 +47,12 @@ description: AI 代理插件配置参考

 #### OpenAI

-OpenAI 所对应的 `type` 为 `openai`。它并无特有的配置字段。
+OpenAI 所对应的 `type` 为 `openai`。它特有的配置字段如下:
+
+| 名称              | 数据类型 | 填写要求 | 默认值 | 描述                                                                          |
+|-------------------|----------|----------|--------|-------------------------------------------------------------------------------|
+| `openaiCustomUrl` | string   | 非必填   | -      | 基于OpenAI协议的自定义后端URL，例如: www.example.com/myai/v1/chat/completions |
+

 #### Azure OpenAI

@@ -89,13 +101,25 @@ DeepSeek所对应的 `type` 为 `deepseek`。它并无特有的配置字段。

 Groq 所对应的 `type` 为 `groq`。它并无特有的配置字段。

+#### 文心一言（Baidu）
+
+文心一言所对应的 `type` 为 `baidu`。它并无特有的配置字段。
+
+#### MiniMax
+
+MiniMax所对应的 `type` 为 `minimax`。它特有的配置字段如下：
+
+| 名称             | 数据类型 | 填写要求                                                     | 默认值 | 描述                                                         |
+| ---------------- | -------- | ------------------------------------------------------------ | ------ | ------------------------------------------------------------ |
+| `minimaxGroupId` | string   | 当使用`abab6.5-chat`, `abab6.5s-chat`, `abab5.5s-chat`, `abab5.5-chat`四种模型时必填 | -      | 当使用`abab6.5-chat`, `abab6.5s-chat`, `abab5.5s-chat`, `abab5.5-chat`四种模型时会使用ChatCompletion Pro，需要设置groupID |
+
 #### Anthropic Claude

 Anthropic Claude 所对应的 `type` 为 `claude`。它特有的配置字段如下：

-| 名称        | 数据类型   | 填写要求 | 默认值 | 描述                |
-|-----------|--------|-----|-----|-------------------|
-| `version` | string | 必填  | -   | Claude 服务的 API 版本 |
+| 名称        | 数据类型   | 填写要求 | 默认值 | 描述                               |
+|-----------|--------|------|-----|----------------------------------|
+| `claudeVersion` | string | 可选   | -   | Claude 服务的 API 版本，默认为 2023-06-01 |

 #### Ollama

@@ -106,6 +130,41 @@ Ollama 所对应的 `type` 为 `ollama`。它特有的配置字段如下：
 | `ollamaServerHost` | string | 必填   | -   | Ollama 服务器的主机地址 |
 | `ollamaServerPort` | number | 必填   | -   | Ollama 服务器的端口号，默认为11434 |

+#### 混元
+
+混元所对应的 `type` 为 `hunyuan`。它特有的配置字段如下：
+
+| 名称                | 数据类型   | 填写要求 | 默认值 | 描述                                           |
+|-------------------|--------|------|-----|----------------------------------------------|
+| `hunyuanAuthId` | string | 必填   | -   | 混元用于v3版本认证的id |
+| `hunyuanAuthKey` | string | 必填   | -   | 混元用于v3版本认证的key |
+
+#### 阶跃星辰 (Stepfun)
+
+阶跃星辰所对应的 `type` 为 `stepfun`。它并无特有的配置字段。
+
+#### Cloudflare Workers AI
+
+Cloudflare Workers AI 所对应的 `type` 为 `cloudflare`。它特有的配置字段如下：
+
+| 名称                | 数据类型   | 填写要求 | 默认值 | 描述                                                                                                                         |
+|-------------------|--------|------|-----|----------------------------------------------------------------------------------------------------------------------------|
+| `cloudflareAccountId` | string | 必填   | -   | [Cloudflare Account ID](https://developers.cloudflare.com/workers-ai/get-started/rest-api/#1-get-api-token-and-account-id) |
+
+#### 星火 (Spark)
+
+星火所对应的 `type` 为 `spark`。它并无特有的配置字段。
+
+讯飞星火认知大模型的`apiTokens`字段值为`APIKey:APISecret`。即填入自己的APIKey与APISecret，并以`:`分隔。
+
+#### Gemini
+
+Gemini 所对应的 `type` 为 `gemini`。它特有的配置字段如下：
+
+| 名称                  | 数据类型 | 填写要求 | 默认值 | 描述                                                                                              |
+| --------------------- | -------- | -------- |-----|-------------------------------------------------------------------------------------------------|
+| `geminiSafetySetting` | map of string   | 非必填     | -   | Gemini AI内容过滤和安全级别设定。参考[Safety settings](https://ai.google.dev/gemini-api/docs/safety-settings) |
+
 ## 用法示例

 ### 使用 OpenAI 协议代理 Azure OpenAI 服务
@@ -221,25 +280,72 @@ provider:
    'gpt-3': "qwen-turbo"
    'gpt-35-turbo': "qwen-plus"
    'gpt-4-turbo': "qwen-max"
+    'gpt-4-*': "qwen-max"
+    'text-embedding-v1': 'text-embedding-v1'
    '*': "qwen-turbo"
 ```

+**AI 对话请求示例**
+
+URL: http://your-domain/v1/chat/completions
+
+请求体：
+
+```json
+{
+  "model": "text-embedding-v1",
+  "input": "Hello"
+}
+```
+
+响应体示例：
+
+```json
+{
+  "object": "list",
+  "data": [
+    {
+      "object": "embedding",
+      "index": 0,
+      "embedding": [
+        -1.0437825918197632,
+        5.208984375,
+        3.0483806133270264,
+        -1.7897135019302368,
+        -2.0107421875,
+        ...,
+        0.8125,
+        -1.1759847402572632,
+        0.8174641728401184,
+        1.0432943105697632,
+        -0.5885213017463684
+      ]
+    }
+  ],
+  "model": "text-embedding-v1",
+  "usage": {
+    "prompt_tokens": 1,
+    "total_tokens": 1
+  }
+}
+```
+
 **请求示例**

+URL: http://your-domain/v1/embeddings
+
+示例请求内容：
+
 ```json
 {
-  "model": "gpt-3",
-  "messages": [
-    {
-      "role": "user",
-      "content": "你好，你是谁？"
-    }
-  ],
-  "temperature": 0.3
+    "model": "text-embedding-v1",
+    "input": [
+        "Hello world!"
+    ]
 }
 ```

-**响应示例**
+示例响应内容：

 ```json
 {
@@ -507,6 +613,7 @@ provider:
  type: claude
  apiTokens:
    - "YOUR_CLAUDE_API_TOKEN"
+  version: "2023-06-01"
 ```

 **请求示例**
@@ -528,27 +635,383 @@ provider:

 ```json
 {
-  "id": "msg_01K8iLH18FGN7Xd9deurwtoD",
-  "type": "message",
-  "role": "assistant",
-  "model": "claude-3-opus-20240229",
-  "stop_sequence": null,
-  "usage": {
-    "input_tokens": 16,
-    "output_tokens": 141
-  },
-  "content": [
+  "id": "msg_01Jt3GzyjuzymnxmZERJguLK",
+  "choices": [
    {
-      "type": "text",
-      "text": "你好!我是Claude,一个由Anthropic公司开发的人工智能助手。我的任务是尽我所能帮助人类,比如回答问题,提供建议和意见,协助完成任务等。我掌握了很多知识,也具备一定的分析和推理能力,但我不是人类,也没有实体的身体。很高兴认识你!如果有什么需要帮助的地方,欢迎随时告诉我。"
+      "index": 0,
+      "message": {
+        "role": "assistant",
+        "content": "您好,我是一个由人工智能公司Anthropic开发的聊天助手。我的名字叫Claude,是一个聪明友善、知识渊博的对话系统。很高兴认识您!我可以就各种话题与您聊天,回答问题,提供建议和帮助。我会尽最大努力给您有帮助的回复。希望我们能有个愉快的交流!"
+      },
+      "finish_reason": "stop"
    }
  ],
-  "stop_reason": "end_turn"
+  "created": 1717385918,
+  "model": "claude-3-opus-20240229",
+  "object": "chat.completion",
+  "usage": {
+    "prompt_tokens": 16,
+    "completion_tokens": 126,
+    "total_tokens": 142
+  }
+}
+```
+### 使用 OpenAI 协议代理混元服务
+
+**配置信息**
+
+```yaml
+provider:
+  type: "hunyuan"
+  hunyuanAuthKey: "<YOUR AUTH KEY>"
+  apiTokens:
+    - ""
+  hunyuanAuthId: "<YOUR AUTH ID>"
+  timeout: 1200000
+  modelMapping:
+    "*": "hunyuan-lite"
+```
+
+**请求示例**
+请求脚本：
+```sh
+
+curl --location 'http://<your higress domain>/v1/chat/completions' \
+--header 'Content-Type:  application/json' \
+--data '{
+  "model": "gpt-3",
+  "messages": [
+    {
+      "role": "system",
+      "content": "你是一个名专业的开发人员！"
+    },
+    {
+      "role": "user",
+      "content": "你好，你是谁？"
+    }
+  ],
+  "temperature": 0.3,
+  "stream": false
+}'
+```
+
+**响应示例**
+
+```json
+{
+    "id": "fd140c3e-0b69-4b19-849b-d354d32a6162",
+    "choices": [
+        {
+            "index": 0,
+            "delta": {
+                "role": "assistant",
+                "content": "你好！我是一名专业的开发人员。"
+            },
+            "finish_reason": "stop"
+        }
+    ],
+    "created": 1717493117,
+    "model": "hunyuan-lite",
+    "object": "chat.completion",
+    "usage": {
+        "prompt_tokens": 15,
+        "completion_tokens": 9,
+        "total_tokens": 24
+    }
+}
+```
+
+### 使用 OpenAI 协议代理百度文心一言服务
+
+**配置信息**
+
+```yaml
+provider:
+  type: baidu
+  apiTokens:
+    - "YOUR_BAIDU_API_TOKEN"
+  modelMapping:
+    'gpt-3': "ERNIE-4.0"
+    '*': "ERNIE-4.0"
+```
+
+**请求示例**
+
+```json
+{
+    "model": "gpt-4-turbo",
+    "messages": [
+        {
+            "role": "user",
+            "content": "你好，你是谁？"
+        }
+    ],
+    "stream": false
+}
+```
+
+**响应示例**
+
+```json
+{
+    "id": "as-e90yfg1pk1",
+    "choices": [
+        {
+            "index": 0,
+            "message": {
+                "role": "assistant",
+                "content": "你好，我是文心一言，英文名是ERNIE Bot。我能够与人对话互动，回答问题，协助创作，高效便捷地帮助人们获取信息、知识和灵感。"
+            },
+            "finish_reason": "stop"
+        }
+    ],
+    "created": 1717251488,
+    "model": "ERNIE-4.0",
+    "object": "chat.completion",
+    "usage": {
+        "prompt_tokens": 4,
+        "completion_tokens": 33,
+        "total_tokens": 37
+    }
+}
+```
+
+### 使用 OpenAI 协议代理MiniMax服务
+
+**配置信息**
+
+```yaml
+provider:
+  type: minimax
+  apiTokens:
+    - "YOUR_MINIMAX_API_TOKEN"
+  modelMapping:
+    "gpt-3": "abab6.5g-chat"
+    "gpt-4": "abab6.5-chat"
+    "*": "abab6.5g-chat"
+  minimaxGroupId: "YOUR_MINIMAX_GROUP_ID"
+```
+
+**请求示例**
+
+```json
+{
+    "model": "gpt-4-turbo",
+    "messages": [
+        {
+            "role": "user",
+            "content": "你好，你是谁？"
+        }
+    ],
+    "stream": false
+}
+```
+
+**响应示例**
+
+```json
+{
+    "id": "02b2251f8c6c09d68c1743f07c72afd7",
+    "choices": [
+        {
+            "finish_reason": "stop",
+            "index": 0,
+            "message": {
+                "content": "你好！我是MM智能助理，一款由MiniMax自研的大型语言模型。我可以帮助你解答问题，提供信息，进行对话等。有什么可以帮助你的吗？",
+                "role": "assistant"
+            }
+        }
+    ],
+    "created": 1717760544,
+    "model": "abab6.5s-chat",
+    "object": "chat.completion",
+    "usage": {
+        "total_tokens": 106
+    },
+    "input_sensitive": false,
+    "output_sensitive": false,
+    "input_sensitive_type": 0,
+    "output_sensitive_type": 0,
+    "base_resp": {
+        "status_code": 0,
+        "status_msg": ""
+    }
+}
+```
+
+### 使用 OpenAI 协议代理 Cloudflare Workers AI 服务
+
+**配置信息**
+
+```yaml
+provider:
+  type: cloudflare
+  apiTokens:
+    - "YOUR_WORKERS_AI_API_TOKEN"
+  cloudflareAccountId: "YOUR_CLOUDFLARE_ACCOUNT_ID"
+  modelMapping:
+    "*": "@cf/meta/llama-3-8b-instruct"
+```
+
+**请求示例**
+
+```json
+{
+  "model": "gpt-3.5",
+  "max_tokens": 1024,
+  "messages": [
+    {
+      "role": "user",
+      "content": "Who are you?"
+    }
+  ]
+}
+```
+
+**响应示例**
+
+```json
+{
+  "id": "id-1720367803430",
+  "object": "chat.completion",
+  "created": 1720367803,
+  "model": "@cf/meta/llama-3-8b-instruct",
+  "choices": [
+    {
+      "index": 0,
+      "message": {
+        "role": "assistant",
+        "content": "I am LLaMA, an AI assistant developed by Meta AI that can understand and respond to human input in a conversational manner. I'm not a human, but a computer program designed to simulate conversation and answer questions to the best of my knowledge. I can be used to generate text on a wide range of topics, from science and history to entertainment and culture.\n\nI'm a large language model, which means I've been trained on a massive dataset of text from the internet and can generate human-like responses. I can understand natural language and respond accordingly, making me suitable for tasks such as:\n\n* Answering questions on various topics\n* Generating text based on a given prompt\n* Translating text from one language to another\n* Summarizing long pieces of text\n* Creating chatbot dialogues\n\nI'm constantly learning and improving, so the more conversations I have with users like you, the better I'll become."
+      },
+      "logprobs": null,
+      "finish_reason": "stop"
+    }
+  ]
+}
+```
+
+### 使用 OpenAI 协议代理Spark服务
+
+**配置信息**
+
+```yaml
+provider:
+  type: spark
+  apiTokens:
+    - "APIKey:APISecret"
+  modelMapping:
+    "gpt-4o": "generalv3.5"
+    "gpt-4": "generalv3"
+    "*": "general"
+```
+
+**请求示例**
+
+```json
+{
+    "model": "gpt-4o",
+    "messages": [
+        {
+            "role": "system",
+            "content": "你是一名专业的开发人员！"
+        },
+        {
+            "role": "user",
+            "content": "你好，你是谁？"
+        }
+    ],
+    "stream": false
+}
+```
+
+**响应示例**
+
+```json
+{
+    "id": "cha000c23c6@dx190ef0b4b96b8f2532",
+    "choices": [
+        {
+            "index": 0,
+            "message": {
+                "role": "assistant",
+                "content": "你好！我是一名专业的开发人员，擅长编程和解决技术问题。有什么我可以帮助你的吗？"
+            }
+        }
+    ],
+    "created": 1721997415,
+    "model": "generalv3.5",
+    "object": "chat.completion",
+    "usage": {
+        "prompt_tokens": 10,
+        "completion_tokens": 19,
+        "total_tokens": 29
+    }
+}
+```
+
+### 使用 OpenAI 协议代理 gemini 服务
+
+**配置信息**
+
+```yaml
+provider:
+  type: gemini
+  apiTokens:
+    - "YOUR_GEMINI_API_TOKEN"
+  modelMapping:
+    "*": "gemini-pro"
+  geminiSafetySetting:
+    "HARM_CATEGORY_SEXUALLY_EXPLICIT" :"BLOCK_NONE"
+    "HARM_CATEGORY_HATE_SPEECH" :"BLOCK_NONE"
+    "HARM_CATEGORY_HARASSMENT" :"BLOCK_NONE"
+    "HARM_CATEGORY_DANGEROUS_CONTENT" :"BLOCK_NONE"
+```
+
+**请求示例**
+
+```json
+{
+    "model": "gpt-3.5",
+    "messages": [
+        {
+            "role": "user",
+            "content": "Who are you?"
+        }
+    ],
+    "stream": false
+}
+```
+
+**响应示例**
+
+```json
+{
+    "id": "chatcmpl-b010867c-0d3f-40ba-95fd-4e8030551aeb",
+    "choices": [
+        {
+            "index": 0,
+            "message": {
+                "role": "assistant",
+                "content": "I am a large multi-modal model, trained by Google. I am designed to provide information and answer questions to the best of my abilities."
+            },
+            "finish_reason": "stop"
+        }
+    ],
+    "created": 1722756984,
+    "model": "gemini-pro",
+    "object": "chat.completion",
+    "usage": {
+        "prompt_tokens": 5,
+        "completion_tokens": 29,
+        "total_tokens": 34
+    }
 }
 ```

 ## 完整配置示例

+### Kubernetes 示例
+
 以下以使用 OpenAI 协议代理 Groq 服务为例，展示完整的插件配置示例。

 ```yaml
@@ -619,4 +1082,131 @@ curl "http://<YOUR-DOMAIN>/v1/chat/completions" -H "Content-Type: application/js
    }
  ]
 }'
-```
+```
+
+### Docker-Compose 示例
+
+`docker-compose.yml` 配置文件：
+
+```yaml
+version: '3.7'
+services:
+  envoy:
+    image: higress-registry.cn-hangzhou.cr.aliyuncs.com/higress/envoy:1.20
+    entrypoint: /usr/local/bin/envoy
+    # 开启了 debug 级别日志方便调试
+    command: -c /etc/envoy/envoy.yaml --component-log-level wasm:debug
+    networks:
+      - higress-net
+    ports:
+      - "10000:10000"
+    volumes:
+      - ./envoy.yaml:/etc/envoy/envoy.yaml
+      - ./plugin.wasm:/etc/envoy/plugin.wasm
+networks:
+  higress-net: {}
+```
+
+`envoy.yaml` 配置文件：
+
+```yaml
+admin:
+  address:
+    socket_address:
+      protocol: TCP
+      address: 0.0.0.0
+      port_value: 9901
+static_resources:
+  listeners:
+    - name: listener_0
+      address:
+        socket_address:
+          protocol: TCP
+          address: 0.0.0.0
+          port_value: 10000
+      filter_chains:
+        - filters:
+            - name: envoy.filters.network.http_connection_manager
+              typed_config:
+                "@type": type.googleapis.com/envoy.extensions.filters.network.http_connection_manager.v3.HttpConnectionManager
+                scheme_header_transformation:
+                  scheme_to_overwrite: https
+                stat_prefix: ingress_http
+                # Output envoy logs to stdout
+                access_log:
+                  - name: envoy.access_loggers.stdout
+                    typed_config:
+                      "@type": type.googleapis.com/envoy.extensions.access_loggers.stream.v3.StdoutAccessLog
+                # Modify as required
+                route_config:
+                  name: local_route
+                  virtual_hosts:
+                    - name: local_service
+                      domains: [ "*" ]
+                      routes:
+                        - match:
+                            prefix: "/"
+                          route:
+                            cluster: claude
+                            timeout: 300s
+                http_filters:
+                  - name: claude
+                    typed_config:
+                      "@type": type.googleapis.com/udpa.type.v1.TypedStruct
+                      type_url: type.googleapis.com/envoy.extensions.filters.http.wasm.v3.Wasm
+                      value:
+                        config:
+                          name: claude
+                          vm_config:
+                            runtime: envoy.wasm.runtime.v8
+                            code:
+                              local:
+                                filename: /etc/envoy/plugin.wasm
+                          configuration:
+                            "@type": "type.googleapis.com/google.protobuf.StringValue"
+                            value: | # 插件配置
+                              {
+                                "provider": {
+                                  "type": "claude",                                
+                                  "apiTokens": [
+                                    "YOUR_API_TOKEN"
+                                  ]                  
+                                }
+                              }
+                  - name: envoy.filters.http.router
+  clusters:
+    - name: claude
+      connect_timeout: 30s
+      type: LOGICAL_DNS
+      dns_lookup_family: V4_ONLY
+      lb_policy: ROUND_ROBIN
+      load_assignment:
+        cluster_name: claude
+        endpoints:
+          - lb_endpoints:
+              - endpoint:
+                  address:
+                    socket_address:
+                      address: api.anthropic.com # API 服务地址
+                      port_value: 443
+      transport_socket:
+        name: envoy.transport_sockets.tls
+        typed_config:
+          "@type": type.googleapis.com/envoy.extensions.transport_sockets.tls.v3.UpstreamTlsContext
+          "sni": "api.anthropic.com"
+```
+
+访问示例：
+
+```bash
+curl "http://localhost:10000/v1/chat/completions"  -H "Content-Type: application/json"  -d '{
+  "model": "claude-3-opus-20240229",
+  "max_tokens": 1024,
+  "messages": [
+    {
+      "role": "user",
+      "content": "你好，你是谁？"
+    }
+  ]
+}'
+```
--- a/plugins/wasm-go/extensions/ai-proxy/README_dev.md
+++ b/plugins/wasm-go/extensions/ai-proxy/README_dev.md
@@ -13,6 +13,52 @@ DOCKER_BUILDKIT=1; docker build --build-arg PLUGIN_NAME=ai-proxy --build-arg EXT
 ```powershell
 $env:DOCKER_BUILDKIT=1; docker build --build-arg PLUGIN_NAME=ai-proxy --build-arg EXTRA_TAGS=proxy_wasm_version_0_2_100 --build-arg BUILDER=higress-registry.cn-hangzhou.cr.aliyuncs.com/plugins/wasm-go-builder:go1.19-tinygo0.28.1-oras1.0.0 -t ai-proxy:0.0.1 --output .\out ..\..
 ```
+
+## 本地运行
+参考：https://higress.io/zh-cn/docs/user/wasm-go
+需要注意的是，higress/plugins/wasm-go/extensions/ai-proxy/envoy.yaml中的clusters字段，记得改成你需要地址，比如混元的话：就会有如下的一个cluster的配置：
+```yaml
+<省略>
+static_resources:
+<省略>
+  clusters:
+      load_assignment:
+        cluster_name: moonshot
+        endpoints:
+          - lb_endpoints:
+              - endpoint:
+                  address:
+                    socket_address:
+                      address: hunyuan.tencentcloudapi.com
+                      port_value: 443
+      transport_socket:
+        name: envoy.transport_sockets.tls
+        typed_config:
+          "@type": type.googleapis.com/envoy.extensions.transport_sockets.tls.v3.UpstreamTlsContext
+          "sni": "hunyuan.tencentcloudapi.com"
+```
+
+而后你就可以在本地的pod中查看相应的输出，请求样例如下：
+```sh
+curl --location 'http://127.0.0.1:10000/v1/chat/completions' \
+--header 'Content-Type:  application/json' \
+--data '{
+  "model": "gpt-3",
+  "messages": [
+    {
+      "role": "system",
+      "content": "你是一个名专业的开发人员！"
+    },
+    {
+      "role": "user",
+      "content": "你好，你是谁？"
+    }
+  ],
+  "temperature": 0.3,
+  "stream": false
+}'
+```
+
 ## 测试须知

 由于 `ai-proxy` 插件使用了 Higress 对数据面定制的特殊功能，因此在测试时需要使用版本不低于 1.4.0-rc.1 的 Higress Gateway 镜像。
--- a/plugins/wasm-go/extensions/ai-proxy/go.mod
+++ b/plugins/wasm-go/extensions/ai-proxy/go.mod
@@ -8,7 +8,7 @@ replace github.com/alibaba/higress/plugins/wasm-go => ../..

 require (
 	github.com/alibaba/higress/plugins/wasm-go v0.0.0
-	github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240327114451-d6b7174a84fc
+	github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240711023527-ba358c48772f
 	github.com/stretchr/testify v1.8.4
 	github.com/tidwall/gjson v1.14.3
 )
--- a/plugins/wasm-go/extensions/ai-proxy/go.sum
+++ b/plugins/wasm-go/extensions/ai-proxy/go.sum
@@ -4,8 +4,8 @@ github.com/google/uuid v1.3.0 h1:t6JiXgmwXMjEs8VusXIJk2BXHsn+wx8BZdTaoZ5fu7I=
 github.com/google/uuid v1.3.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
 github.com/higress-group/nottinygc v0.0.0-20231101025119-e93c4c2f8520 h1:IHDghbGQ2DTIXHBHxWfqCYQW1fKjyJ/I7W1pMyUDeEA=
 github.com/higress-group/nottinygc v0.0.0-20231101025119-e93c4c2f8520/go.mod h1:Nz8ORLaFiLWotg6GeKlJMhv8cci8mM43uEnLA5t8iew=
-github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240327114451-d6b7174a84fc h1:t2AT8zb6N/59Y78lyRWedVoVWHNRSCBh0oWCC+bluTQ=
-github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240327114451-d6b7174a84fc/go.mod h1:hNFjhrLUIq+kJ9bOcs8QtiplSQ61GZXtd2xHKx4BYRo=
+github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240711023527-ba358c48772f h1:ZIiIBRvIw62gA5MJhuwp1+2wWbqL9IGElQ499rUsYYg=
+github.com/higress-group/proxy-wasm-go-sdk v0.0.0-20240711023527-ba358c48772f/go.mod h1:hNFjhrLUIq+kJ9bOcs8QtiplSQ61GZXtd2xHKx4BYRo=
 github.com/magefile/mage v1.14.0 h1:6QDX3g6z1YvJ4olPhT1wksUcSa/V0a1B+pJb73fBjyo=
 github.com/magefile/mage v1.14.0/go.mod h1:z5UZb/iS3GoOSn0JgWuiw7dxlurVYTu+/jHXqQg881A=
 github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
--- a/plugins/wasm-go/extensions/ai-proxy/main.go
+++ b/plugins/wasm-go/extensions/ai-proxy/main.go
@@ -21,6 +21,8 @@ const (
 	pluginName = "ai-proxy"

 	ctxKeyApiName = "apiKey"
+
+	defaultMaxBodyBytes uint32 = 10 * 1024 * 1024
 )

 func main() {
@@ -61,10 +63,10 @@ func onHttpRequestHeader(ctx wrapper.HttpContext, pluginConfig config.PluginConf

 	rawPath := ctx.Path()
 	path, _ := url.Parse(rawPath)
-	apiName := getApiName(path.Path)
+	apiName := getOpenAiApiName(path.Path)
 	if apiName == "" {
 		log.Debugf("[onHttpRequestHeader] unsupported path: %s", path.Path)
-		_ = util.SendResponse(404, util.MimeTypeTextPlain, "API not found: "+path.Path)
+		_ = util.SendResponse(404, "ai-proxy.unknown_api", util.MimeTypeTextPlain, "API not found: "+path.Path)
 		return types.ActionContinue
 	}
 	ctx.SetContext(ctxKeyApiName, apiName)
@@ -75,16 +77,18 @@ func onHttpRequestHeader(ctx wrapper.HttpContext, pluginConfig config.PluginConf

 		action, err := handler.OnRequestHeaders(ctx, apiName, log)
 		if err == nil {
+			if contentType, err := proxywasm.GetHttpRequestHeader("Content-Type"); err == nil && contentType != "" {
+				ctx.SetRequestBodyBufferLimit(defaultMaxBodyBytes)
+				// Always return types.HeaderStopIteration to support fallback routing,
+				// as long as onHttpRequestBody can be called.
+				return types.HeaderStopIteration
+			}
 			return action
 		}
-		_ = util.SendResponse(404, util.MimeTypeTextPlain, fmt.Sprintf("failed to process request headers: %v", err))
+		_ = util.SendResponse(500, "ai-proxy.proc_req_headers_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to process request headers: %v", err))
 		return types.ActionContinue
 	}

-	if _, needHandleBody := activeProvider.(provider.RequestBodyHandler); needHandleBody {
-		ctx.DontReadRequestBody()
-	}
-
 	return types.ActionContinue
 }

@@ -99,18 +103,24 @@ func onHttpRequestBody(ctx wrapper.HttpContext, pluginConfig config.PluginConfig
 	log.Debugf("[onHttpRequestBody] provider=%s", activeProvider.GetProviderType())

 	if handler, ok := activeProvider.(provider.RequestBodyHandler); ok {
-		apiName := ctx.GetContext(ctxKeyApiName).(provider.ApiName)
+		apiName, _ := ctx.GetContext(ctxKeyApiName).(provider.ApiName)
 		action, err := handler.OnRequestBody(ctx, apiName, body, log)
 		if err == nil {
 			return action
 		}
-		_ = util.SendResponse(404, util.MimeTypeTextPlain, fmt.Sprintf("failed to process request body: %v", err))
+		_ = util.SendResponse(500, "ai-proxy.proc_req_body_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to process request body: %v", err))
 		return types.ActionContinue
 	}
 	return types.ActionContinue
 }

 func onHttpResponseHeaders(ctx wrapper.HttpContext, pluginConfig config.PluginConfig, log wrapper.Log) types.Action {
+	if !wrapper.IsResponseFromUpstream() {
+		// Response is not coming from the upstream. Let it pass through.
+		ctx.DontReadResponseBody()
+		return types.ActionContinue
+	}
+
 	activeProvider := pluginConfig.GetProvider()

 	if activeProvider == nil {
@@ -139,12 +149,12 @@ func onHttpResponseHeaders(ctx wrapper.HttpContext, pluginConfig config.PluginCo
 	}

 	if handler, ok := activeProvider.(provider.ResponseHeadersHandler); ok {
-		apiName := ctx.GetContext(ctxKeyApiName).(provider.ApiName)
+		apiName, _ := ctx.GetContext(ctxKeyApiName).(provider.ApiName)
 		action, err := handler.OnResponseHeaders(ctx, apiName, log)
 		if err == nil {
 			return action
 		}
-		_ = util.SendResponse(404, util.MimeTypeTextPlain, fmt.Sprintf("failed to process response headers: %v", err))
+		_ = util.SendResponse(500, "ai-proxy.proc_resp_headers_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to process response headers: %v", err))
 		return types.ActionContinue
 	}

@@ -171,7 +181,7 @@ func onStreamingResponseBody(ctx wrapper.HttpContext, pluginConfig config.Plugin
 	log.Debugf("isLastChunk=%v chunk: %s", isLastChunk, string(chunk))

 	if handler, ok := activeProvider.(provider.StreamingResponseBodyHandler); ok {
-		apiName := ctx.GetContext(ctxKeyApiName).(provider.ApiName)
+		apiName, _ := ctx.GetContext(ctxKeyApiName).(provider.ApiName)
 		modifiedChunk, err := handler.OnStreamingResponseBody(ctx, apiName, chunk, isLastChunk, log)
 		if err == nil && modifiedChunk != nil {
 			return modifiedChunk
@@ -193,20 +203,23 @@ func onHttpResponseBody(ctx wrapper.HttpContext, pluginConfig config.PluginConfi
 	//log.Debugf("response body: %s", string(body))

 	if handler, ok := activeProvider.(provider.ResponseBodyHandler); ok {
-		apiName := ctx.GetContext(ctxKeyApiName).(provider.ApiName)
+		apiName, _ := ctx.GetContext(ctxKeyApiName).(provider.ApiName)
 		action, err := handler.OnResponseBody(ctx, apiName, body, log)
 		if err == nil {
 			return action
 		}
-		_ = util.SendResponse(404, util.MimeTypeTextPlain, fmt.Sprintf("failed to process response body: %v", err))
+		_ = util.SendResponse(500, "ai-proxy.proc_resp_body_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to process response body: %v", err))
 		return types.ActionContinue
 	}
 	return types.ActionContinue
 }

-func getApiName(path string) provider.ApiName {
+func getOpenAiApiName(path string) provider.ApiName {
 	if strings.HasSuffix(path, "/v1/chat/completions") {
 		return provider.ApiNameChatCompletion
 	}
+	if strings.HasSuffix(path, "/v1/embeddings") {
+		return provider.ApiNameEmbeddings
+	}
 	return ""
 }
--- a/plugins/wasm-go/extensions/ai-proxy/provider/azure.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/azure.go
@@ -23,6 +23,9 @@ func (m *azureProviderInitializer) ValidateConfig(config ProviderConfig) error {
 	if _, err := url.Parse(config.azureServiceUrl); err != nil {
 		return fmt.Errorf("invalid azureServiceUrl: %w", err)
 	}
+	if config.apiTokens == nil || len(config.apiTokens) == 0 {
+		return errors.New("no apiToken found in provider config")
+	}
 	return nil
 }

@@ -52,44 +55,43 @@ func (m *azureProvider) GetProviderType() string {
 }

 func (m *azureProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
-	if apiName != ApiNameChatCompletion {
-		return types.ActionContinue, errUnsupportedApiName
-	}
 	_ = util.OverwriteRequestPath(m.serviceUrl.RequestURI())
 	_ = util.OverwriteRequestHost(m.serviceUrl.Host)
 	_ = proxywasm.ReplaceHttpRequestHeader("api-key", m.config.apiTokens[0])
-
-	if m.contextCache == nil {
-		ctx.DontReadRequestBody()
-	} else {
+	if apiName == ApiNameChatCompletion {
 		_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
+	} else {
+		ctx.DontReadRequestBody()
 	}
-
 	return types.ActionContinue, nil
 }

 func (m *azureProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
 	if apiName != ApiNameChatCompletion {
-		return types.ActionContinue, errUnsupportedApiName
-	}
-	if m.contextCache == nil {
+		// We don't need to process the request body for other APIs.
 		return types.ActionContinue, nil
 	}
 	request := &chatCompletionRequest{}
 	if err := decodeChatCompletionRequest(body, request); err != nil {
 		return types.ActionContinue, err
 	}
+	if m.contextCache == nil {
+		if err := replaceJsonRequestBody(request, log); err != nil {
+			_ = util.SendResponse(500, "ai-proxy.openai.set_include_usage_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+		}
+		return types.ActionContinue, nil
+	}
 	err := m.contextCache.GetContent(func(content string, err error) {
 		defer func() {
 			_ = proxywasm.ResumeHttpRequest()
 		}()
 		if err != nil {
 			log.Errorf("failed to load context file: %v", err)
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.azure.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
 		}
 		insertContextMessage(request, content)
 		if err := replaceJsonRequestBody(request, log); err != nil {
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.azure.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
 		}
 	}, log)
 	if err == nil {
--- a/plugins/wasm-go/extensions/ai-proxy/provider/baichuan.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/baichuan.go
@@ -1,6 +1,7 @@
 package provider

 import (
+	"errors"
 	"fmt"

 	"github.com/alibaba/higress/plugins/wasm-go/extensions/ai-proxy/util"
@@ -20,6 +21,9 @@ type baichuanProviderInitializer struct {
 }

 func (m *baichuanProviderInitializer) ValidateConfig(config ProviderConfig) error {
+	if config.apiTokens == nil || len(config.apiTokens) == 0 {
+		return errors.New("no apiToken found in provider config")
+	}
 	return nil
 }

@@ -45,14 +49,8 @@ func (m *baichuanProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName Api
 	}
 	_ = util.OverwriteRequestPath(baichuanChatCompletionPath)
 	_ = util.OverwriteRequestHost(baichuanDomain)
-	_ = proxywasm.ReplaceHttpRequestHeader("Authorization", "Bearer "+m.config.GetRandomToken())
-
-	if m.contextCache == nil {
-		ctx.DontReadRequestBody()
-	} else {
-		_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
-	}
-
+	_ = util.OverwriteRequestAuthorization("Bearer " + m.config.GetRandomToken())
+	_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
 	return types.ActionContinue, nil
 }

@@ -73,11 +71,11 @@ func (m *baichuanProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiNam
 		}()
 		if err != nil {
 			log.Errorf("failed to load context file: %v", err)
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.baichuan.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
 		}
 		insertContextMessage(request, content)
 		if err := replaceJsonRequestBody(request, log); err != nil {
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.baichuan.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
 		}
 	}, log)
 	if err == nil {
--- a/plugins/wasm-go/extensions/ai-proxy/provider/baidu.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/baidu.go
@@ -0,0 +1,341 @@
+package provider
+
+import (
+	"encoding/json"
+	"errors"
+	"fmt"
+	"strings"
+	"time"
+
+	"github.com/alibaba/higress/plugins/wasm-go/extensions/ai-proxy/util"
+	"github.com/alibaba/higress/plugins/wasm-go/pkg/wrapper"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm/types"
+)
+
+// baiduProvider is the provider for baidu ernie bot service.
+
+const (
+	baiduDomain = "aip.baidubce.com"
+)
+
+var baiduModelToPathSuffixMap = map[string]string{
+	"ERNIE-4.0-8K":     "completions_pro",
+	"ERNIE-3.5-8K":     "completions",
+	"ERNIE-3.5-128K":   "ernie-3.5-128k",
+	"ERNIE-Speed-8K":   "ernie_speed",
+	"ERNIE-Speed-128K": "ernie-speed-128k",
+	"ERNIE-Tiny-8K":    "ernie-tiny-8k",
+	"ERNIE-Bot-8K":     "ernie_bot_8k",
+	"BLOOMZ-7B":        "bloomz_7b1",
+}
+
+type baiduProviderInitializer struct {
+}
+
+func (b *baiduProviderInitializer) ValidateConfig(config ProviderConfig) error {
+	if config.apiTokens == nil || len(config.apiTokens) == 0 {
+		return errors.New("no apiToken found in provider config")
+	}
+	return nil
+}
+
+func (b *baiduProviderInitializer) CreateProvider(config ProviderConfig) (Provider, error) {
+	return &baiduProvider{
+		config:       config,
+		contextCache: createContextCache(&config),
+	}, nil
+}
+
+type baiduProvider struct {
+	config       ProviderConfig
+	contextCache *contextCache
+}
+
+func (b *baiduProvider) GetProviderType() string {
+	return providerTypeBaidu
+}
+
+func (b *baiduProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+	_ = util.OverwriteRequestHost(baiduDomain)
+
+	_ = proxywasm.RemoveHttpRequestHeader("Accept-Encoding")
+	_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
+
+	// Delay the header processing to allow changing streaming mode in OnRequestBody
+	return types.HeaderStopIteration, nil
+}
+
+func (b *baiduProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+	// 使用文心一言接口协议
+	if b.config.protocol == protocolOriginal {
+		request := &baiduTextGenRequest{}
+		if err := json.Unmarshal(body, request); err != nil {
+			return types.ActionContinue, fmt.Errorf("unable to unmarshal request: %v", err)
+		}
+		if request.Model == "" {
+			return types.ActionContinue, errors.New("request model is empty")
+		}
+		// 根据模型重写requestPath
+		path := b.getRequestPath(request.Model)
+		_ = util.OverwriteRequestPath(path)
+
+		if b.config.context == nil {
+			return types.ActionContinue, nil
+		}
+
+		err := b.contextCache.GetContent(func(content string, err error) {
+			defer func() {
+				_ = proxywasm.ResumeHttpRequest()
+			}()
+
+			if err != nil {
+				log.Errorf("failed to load context file: %v", err)
+				_ = util.SendResponse(500, "ai-proxy.baidu.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+			}
+			b.setSystemContent(request, content)
+			if err := replaceJsonRequestBody(request, log); err != nil {
+				_ = util.SendResponse(500, "ai-proxy.baidu.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+			}
+		}, log)
+		if err == nil {
+			return types.ActionPause, nil
+		}
+		return types.ActionContinue, err
+	}
+	request := &chatCompletionRequest{}
+	if err := decodeChatCompletionRequest(body, request); err != nil {
+		return types.ActionContinue, err
+	}
+
+	// 映射模型重写requestPath
+	model := request.Model
+	if model == "" {
+		return types.ActionContinue, errors.New("missing model in chat completion request")
+	}
+	ctx.SetContext(ctxKeyOriginalRequestModel, model)
+	mappedModel := getMappedModel(model, b.config.modelMapping, log)
+	if mappedModel == "" {
+		return types.ActionContinue, errors.New("model becomes empty after applying the configured mapping")
+	}
+	request.Model = mappedModel
+	ctx.SetContext(ctxKeyFinalRequestModel, request.Model)
+	path := b.getRequestPath(mappedModel)
+	_ = util.OverwriteRequestPath(path)
+
+	if b.config.context == nil {
+		baiduRequest := b.baiduTextGenRequest(request)
+		return types.ActionContinue, replaceJsonRequestBody(baiduRequest, log)
+	}
+
+	err := b.contextCache.GetContent(func(content string, err error) {
+		defer func() {
+			_ = proxywasm.ResumeHttpRequest()
+		}()
+		if err != nil {
+			log.Errorf("failed to load context file: %v", err)
+			_ = util.SendResponse(500, "ai-proxy.baidu.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+		}
+		insertContextMessage(request, content)
+		baiduRequest := b.baiduTextGenRequest(request)
+		if err := replaceJsonRequestBody(baiduRequest, log); err != nil {
+			_ = util.SendResponse(500, "ai-proxy.baidu.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace Request body: %v", err))
+		}
+	}, log)
+	if err == nil {
+		return types.ActionPause, nil
+	}
+	return types.ActionContinue, err
+}
+
+func (b *baiduProvider) OnResponseHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	// 使用文心一言接口协议,跳过OnStreamingResponseBody()和OnResponseBody()
+	if b.config.protocol == protocolOriginal {
+		ctx.DontReadResponseBody()
+		return types.ActionContinue, nil
+	}
+
+	_ = proxywasm.RemoveHttpResponseHeader("Content-Length")
+	return types.ActionContinue, nil
+}
+
+func (b *baiduProvider) OnStreamingResponseBody(ctx wrapper.HttpContext, name ApiName, chunk []byte, isLastChunk bool, log wrapper.Log) ([]byte, error) {
+	if isLastChunk || len(chunk) == 0 {
+		return nil, nil
+	}
+	// sample event response:
+	// data: {"id":"as-vb0m37ti8y","object":"chat.completion","created":1709089502,"sentence_id":0,"is_end":false,"is_truncated":false,"result":"当然可以，","need_clear_history":false,"finish_reason":"normal","usage":{"prompt_tokens":5,"completion_tokens":2,"total_tokens":7}}
+
+	// sample end event response:
+	// data: {"id":"as-vb0m37ti8y","object":"chat.completion","created":1709089531,"sentence_id":20,"is_end":true,"is_truncated":false,"result":"","need_clear_history":false,"finish_reason":"normal","usage":{"prompt_tokens":5,"completion_tokens":420,"total_tokens":425}}
+	responseBuilder := &strings.Builder{}
+	lines := strings.Split(string(chunk), "\n")
+	for _, data := range lines {
+		if len(data) < 6 {
+			// ignore blank line or wrong format
+			continue
+		}
+		data = data[6:]
+		var baiduResponse baiduTextGenStreamResponse
+		if err := json.Unmarshal([]byte(data), &baiduResponse); err != nil {
+			log.Errorf("unable to unmarshal baidu response: %v", err)
+			continue
+		}
+		response := b.streamResponseBaidu2OpenAI(ctx, &baiduResponse)
+		responseBody, err := json.Marshal(response)
+		if err != nil {
+			log.Errorf("unable to marshal response: %v", err)
+			return nil, err
+		}
+		b.appendResponse(responseBuilder, string(responseBody))
+	}
+	modifiedResponseChunk := responseBuilder.String()
+	log.Debugf("=== modified response chunk: %s", modifiedResponseChunk)
+	return []byte(modifiedResponseChunk), nil
+}
+
+func (b *baiduProvider) OnResponseBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	baiduResponse := &baiduTextGenResponse{}
+	if err := json.Unmarshal(body, baiduResponse); err != nil {
+		return types.ActionContinue, fmt.Errorf("unable to unmarshal baidu response: %v", err)
+	}
+	if baiduResponse.ErrorMsg != "" {
+		return types.ActionContinue, fmt.Errorf("baidu response error, error_code: %d, error_message: %s", baiduResponse.ErrorCode, baiduResponse.ErrorMsg)
+	}
+	response := b.responseBaidu2OpenAI(ctx, baiduResponse)
+	return types.ActionContinue, replaceJsonResponseBody(response, log)
+}
+
+type baiduTextGenRequest struct {
+	Model           string        `json:"model"`
+	Messages        []chatMessage `json:"messages"`
+	Temperature     float64       `json:"temperature,omitempty"`
+	TopP            float64       `json:"top_p,omitempty"`
+	PenaltyScore    float64       `json:"penalty_score,omitempty"`
+	Stream          bool          `json:"stream,omitempty"`
+	System          string        `json:"system,omitempty"`
+	DisableSearch   bool          `json:"disable_search,omitempty"`
+	EnableCitation  bool          `json:"enable_citation,omitempty"`
+	MaxOutputTokens int           `json:"max_output_tokens,omitempty"`
+	UserId          string        `json:"user_id,omitempty"`
+}
+
+func (b *baiduProvider) getRequestPath(baiduModel string) string {
+	// https://cloud.baidu.com/doc/WENXINWORKSHOP/s/clntwmv7t
+	suffix, ok := baiduModelToPathSuffixMap[baiduModel]
+	if !ok {
+		suffix = baiduModel
+	}
+	return fmt.Sprintf("/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/%s?access_token=%s", suffix, b.config.GetRandomToken())
+}
+
+func (b *baiduProvider) setSystemContent(request *baiduTextGenRequest, content string) {
+	request.System = content
+}
+
+func (b *baiduProvider) baiduTextGenRequest(request *chatCompletionRequest) *baiduTextGenRequest {
+	baiduRequest := baiduTextGenRequest{
+		Messages:        make([]chatMessage, 0, len(request.Messages)),
+		Temperature:     request.Temperature,
+		TopP:            request.TopP,
+		PenaltyScore:    request.FrequencyPenalty,
+		Stream:          request.Stream,
+		DisableSearch:   false,
+		EnableCitation:  false,
+		MaxOutputTokens: request.MaxTokens,
+		UserId:          request.User,
+	}
+	for _, message := range request.Messages {
+		if message.Role == roleSystem {
+			baiduRequest.System = message.Content
+		} else {
+			baiduRequest.Messages = append(baiduRequest.Messages, chatMessage{
+				Role:    message.Role,
+				Content: message.Content,
+			})
+		}
+	}
+	return &baiduRequest
+}
+
+type baiduTextGenResponse struct {
+	Id               string                    `json:"id"`
+	Object           string                    `json:"object"`
+	Created          int64                     `json:"created"`
+	Result           string                    `json:"result"`
+	IsTruncated      bool                      `json:"is_truncated"`
+	NeedClearHistory bool                      `json:"need_clear_history"`
+	Usage            baiduTextGenResponseUsage `json:"usage"`
+	baiduTextGenResponseError
+}
+
+type baiduTextGenResponseError struct {
+	ErrorCode int    `json:"error_code"`
+	ErrorMsg  string `json:"error_msg"`
+}
+
+type baiduTextGenStreamResponse struct {
+	baiduTextGenResponse
+	SentenceId int  `json:"sentence_id"`
+	IsEnd      bool `json:"is_end"`
+}
+
+type baiduTextGenResponseUsage struct {
+	PromptTokens     int `json:"prompt_tokens"`
+	CompletionTokens int `json:"completion_tokens"`
+	TotalTokens      int `json:"total_tokens"`
+}
+
+func (b *baiduProvider) responseBaidu2OpenAI(ctx wrapper.HttpContext, response *baiduTextGenResponse) *chatCompletionResponse {
+	choice := chatCompletionChoice{
+		Index:        0,
+		Message:      &chatMessage{Role: roleAssistant, Content: response.Result},
+		FinishReason: finishReasonStop,
+	}
+	return &chatCompletionResponse{
+		Id:                response.Id,
+		Created:           time.Now().UnixMilli() / 1000,
+		Model:             ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
+		SystemFingerprint: "",
+		Object:            objectChatCompletion,
+		Choices:           []chatCompletionChoice{choice},
+		Usage: usage{
+			PromptTokens:     response.Usage.PromptTokens,
+			CompletionTokens: response.Usage.CompletionTokens,
+			TotalTokens:      response.Usage.TotalTokens,
+		},
+	}
+}
+
+func (b *baiduProvider) streamResponseBaidu2OpenAI(ctx wrapper.HttpContext, response *baiduTextGenStreamResponse) *chatCompletionResponse {
+	choice := chatCompletionChoice{
+		Index:   0,
+		Message: &chatMessage{Role: roleAssistant, Content: response.Result},
+	}
+	if response.IsEnd {
+		choice.FinishReason = finishReasonStop
+	}
+	return &chatCompletionResponse{
+		Id:                response.Id,
+		Created:           time.Now().UnixMilli() / 1000,
+		Model:             ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
+		SystemFingerprint: "",
+		Object:            objectChatCompletionChunk,
+		Choices:           []chatCompletionChoice{choice},
+		Usage: usage{
+			PromptTokens:     response.Usage.PromptTokens,
+			CompletionTokens: response.Usage.CompletionTokens,
+			TotalTokens:      response.Usage.TotalTokens,
+		},
+	}
+}
+
+func (b *baiduProvider) appendResponse(responseBuilder *strings.Builder, responseBody string) {
+	responseBuilder.WriteString(fmt.Sprintf("%s %s\n\n", streamDataItemKey, responseBody))
+}
--- a/plugins/wasm-go/extensions/ai-proxy/provider/claude.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/claude.go
@@ -0,0 +1,371 @@
+package provider
+
+import (
+	"encoding/json"
+	"errors"
+	"fmt"
+	"strings"
+	"time"
+
+	"github.com/alibaba/higress/plugins/wasm-go/extensions/ai-proxy/util"
+	"github.com/alibaba/higress/plugins/wasm-go/pkg/wrapper"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm/types"
+)
+
+// claudeProvider is the provider for Claude service.
+const (
+	claudeDomain             = "api.anthropic.com"
+	claudeChatCompletionPath = "/v1/messages"
+	defaultVersion           = "2023-06-01"
+	defaultMaxTokens         = 4096
+)
+
+type claudeProviderInitializer struct{}
+
+type claudeTextGenRequest struct {
+	Model         string        `json:"model"`
+	Messages      []chatMessage `json:"messages"`
+	System        string        `json:"system,omitempty"`
+	MaxTokens     int           `json:"max_tokens,omitempty"`
+	StopSequences []string      `json:"stop_sequences,omitempty"`
+	Stream        bool          `json:"stream,omitempty"`
+	Temperature   float64       `json:"temperature,omitempty"`
+	TopP          float64       `json:"top_p,omitempty"`
+	TopK          int           `json:"top_k,omitempty"`
+}
+
+type claudeTextGenResponse struct {
+	Id           string                 `json:"id"`
+	Type         string                 `json:"type"`
+	Role         string                 `json:"role"`
+	Content      []claudeTextGenContent `json:"content"`
+	Model        string                 `json:"model"`
+	StopReason   *string                `json:"stop_reason"`
+	StopSequence *string                `json:"stop_sequence"`
+	Usage        claudeTextGenUsage     `json:"usage"`
+	Error        *claudeTextGenError    `json:"error"`
+}
+
+type claudeTextGenContent struct {
+	Type string `json:"type"`
+	Text string `json:"text,omitempty"`
+}
+
+type claudeTextGenUsage struct {
+	InputTokens  int `json:"input_tokens"`
+	OutputTokens int `json:"output_tokens"`
+}
+
+type claudeTextGenError struct {
+	Type    string `json:"type"`
+	Message string `json:"message"`
+}
+
+type claudeTextGenStreamResponse struct {
+	Type         string                `json:"type"`
+	Message      claudeTextGenResponse `json:"message"`
+	Index        int                   `json:"index"`
+	ContentBlock *claudeTextGenContent `json:"content_block"`
+	Delta        *claudeTextGenDelta   `json:"delta"`
+	Usage        claudeTextGenUsage    `json:"usage"`
+}
+
+type claudeTextGenDelta struct {
+	Type         string  `json:"type"`
+	Text         string  `json:"text"`
+	StopReason   *string `json:"stop_reason"`
+	StopSequence *string `json:"stop_sequence"`
+}
+
+func (c *claudeProviderInitializer) ValidateConfig(config ProviderConfig) error {
+	if config.apiTokens == nil || len(config.apiTokens) == 0 {
+		return errors.New("no apiToken found in provider config")
+	}
+	return nil
+}
+
+func (c *claudeProviderInitializer) CreateProvider(config ProviderConfig) (Provider, error) {
+	return &claudeProvider{
+		config:       config,
+		contextCache: createContextCache(&config),
+	}, nil
+}
+
+type claudeProvider struct {
+	config       ProviderConfig
+	contextCache *contextCache
+}
+
+func (c *claudeProvider) GetProviderType() string {
+	return providerTypeClaude
+}
+
+func (c *claudeProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+
+	_ = util.OverwriteRequestPath(claudeChatCompletionPath)
+	_ = util.OverwriteRequestHost(claudeDomain)
+	_ = proxywasm.ReplaceHttpRequestHeader("x-api-key", c.config.GetRandomToken())
+
+	if c.config.claudeVersion == "" {
+		c.config.claudeVersion = defaultVersion
+	}
+	_ = proxywasm.AddHttpRequestHeader("anthropic-version", c.config.claudeVersion)
+	_ = proxywasm.RemoveHttpRequestHeader("Accept-Encoding")
+	_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
+
+	return types.ActionContinue, nil
+}
+
+func (c *claudeProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+
+	// use original protocol
+	if c.config.protocol == protocolOriginal {
+		if c.config.context == nil {
+			return types.ActionContinue, nil
+		}
+
+		request := &claudeTextGenRequest{}
+		if err := json.Unmarshal(body, request); err != nil {
+			return types.ActionContinue, fmt.Errorf("unable to unmarshal request: %v", err)
+		}
+
+		err := c.contextCache.GetContent(func(content string, err error) {
+			defer func() {
+				_ = proxywasm.ResumeHttpRequest()
+			}()
+
+			if err != nil {
+				log.Errorf("failed to load context file: %v", err)
+				_ = util.SendResponse(500, "ai-proxy.claude.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+			}
+			if err := replaceJsonRequestBody(request, log); err != nil {
+				_ = util.SendResponse(500, "ai-proxy.claude.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+			}
+		}, log)
+		if err == nil {
+			return types.ActionPause, nil
+		}
+		return types.ActionContinue, err
+	}
+
+	// use openai protocol
+	request := &chatCompletionRequest{}
+	if err := decodeChatCompletionRequest(body, request); err != nil {
+		return types.ActionContinue, err
+	}
+
+	model := request.Model
+	if model == "" {
+		return types.ActionContinue, errors.New("missing model in chat completion request")
+	}
+	ctx.SetContext(ctxKeyOriginalRequestModel, model)
+	mappedModel := getMappedModel(model, c.config.modelMapping, log)
+	if mappedModel == "" {
+		return types.ActionContinue, errors.New("model becomes empty after applying the configured mapping")
+	}
+	request.Model = mappedModel
+	ctx.SetContext(ctxKeyFinalRequestModel, request.Model)
+
+	streaming := request.Stream
+	if streaming {
+		_ = proxywasm.ReplaceHttpRequestHeader("Accept", "text/event-stream")
+	}
+
+	if c.config.context == nil {
+		claudeRequest := c.buildClaudeTextGenRequest(request)
+		return types.ActionContinue, replaceJsonRequestBody(claudeRequest, log)
+	}
+
+	err := c.contextCache.GetContent(func(content string, err error) {
+		defer func() {
+			_ = proxywasm.ResumeHttpRequest()
+		}()
+		if err != nil {
+			log.Errorf("failed to load context file: %v", err)
+			_ = util.SendResponse(500, "ai-proxy.claude.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+		}
+		insertContextMessage(request, content)
+		claudeRequest := c.buildClaudeTextGenRequest(request)
+		if err := replaceJsonRequestBody(claudeRequest, log); err != nil {
+			_ = util.SendResponse(500, "ai-proxy.claude.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+		}
+	}, log)
+	if err == nil {
+		return types.ActionPause, nil
+	}
+	return types.ActionContinue, err
+}
+
+func (c *claudeProvider) OnResponseBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	claudeResponse := &claudeTextGenResponse{}
+	if err := json.Unmarshal(body, claudeResponse); err != nil {
+		return types.ActionContinue, fmt.Errorf("unable to unmarshal claude response: %v", err)
+	}
+	if claudeResponse.Error != nil {
+		return types.ActionContinue, fmt.Errorf("claude response error, error_type: %s, error_message: %s", claudeResponse.Error.Type, claudeResponse.Error.Message)
+	}
+	response := c.responseClaude2OpenAI(ctx, claudeResponse)
+	return types.ActionContinue, replaceJsonResponseBody(response, log)
+}
+
+func (c *claudeProvider) OnResponseHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	// use original protocol, skip OnStreamingResponseBody() and OnResponseBody()
+	if c.config.protocol == protocolOriginal {
+		ctx.DontReadResponseBody()
+		return types.ActionContinue, nil
+	}
+
+	_ = proxywasm.RemoveHttpResponseHeader("Content-Length")
+	return types.ActionContinue, nil
+}
+
+func (c *claudeProvider) OnStreamingResponseBody(ctx wrapper.HttpContext, name ApiName, chunk []byte, isLastChunk bool, log wrapper.Log) ([]byte, error) {
+	if isLastChunk || len(chunk) == 0 {
+		return nil, nil
+	}
+
+	responseBuilder := &strings.Builder{}
+	lines := strings.Split(string(chunk), "\n")
+	for _, data := range lines {
+		// only process the line starting with "data:"
+		if strings.HasPrefix(data, "data:") {
+			// extract json data from the line
+			jsonData := strings.TrimPrefix(data, "data:")
+			var claudeResponse claudeTextGenStreamResponse
+			if err := json.Unmarshal([]byte(jsonData), &claudeResponse); err != nil {
+				log.Errorf("unable to unmarshal claude response: %v", err)
+				continue
+			}
+			response := c.streamResponseClaude2OpenAI(ctx, &claudeResponse, log)
+			if response != nil {
+				responseBody, err := json.Marshal(response)
+				if err != nil {
+					log.Errorf("unable to marshal response: %v", err)
+					return nil, err
+				}
+				c.appendResponse(responseBuilder, string(responseBody))
+			}
+		}
+	}
+	modifiedResponseChunk := responseBuilder.String()
+	log.Debugf("modified response chunk: %s", modifiedResponseChunk)
+	return []byte(modifiedResponseChunk), nil
+}
+
+func (c *claudeProvider) buildClaudeTextGenRequest(origRequest *chatCompletionRequest) *claudeTextGenRequest {
+	claudeRequest := claudeTextGenRequest{
+		Model:         origRequest.Model,
+		MaxTokens:     origRequest.MaxTokens,
+		StopSequences: origRequest.Stop,
+		Stream:        origRequest.Stream,
+		Temperature:   origRequest.Temperature,
+		TopP:          origRequest.TopP,
+	}
+	if claudeRequest.MaxTokens == 0 {
+		claudeRequest.MaxTokens = defaultMaxTokens
+	}
+
+	for _, message := range origRequest.Messages {
+		if message.Role == roleSystem {
+			claudeRequest.System = message.Content
+			continue
+		}
+		claudeMessage := chatMessage{
+			Role:    message.Role,
+			Content: message.Content,
+		}
+		claudeRequest.Messages = append(claudeRequest.Messages, claudeMessage)
+	}
+	return &claudeRequest
+}
+
+func (c *claudeProvider) responseClaude2OpenAI(ctx wrapper.HttpContext, origResponse *claudeTextGenResponse) *chatCompletionResponse {
+	choice := chatCompletionChoice{
+		Index:        0,
+		Message:      &chatMessage{Role: roleAssistant, Content: origResponse.Content[0].Text},
+		FinishReason: stopReasonClaude2OpenAI(origResponse.StopReason),
+	}
+
+	return &chatCompletionResponse{
+		Id:                origResponse.Id,
+		Created:           time.Now().UnixMilli() / 1000,
+		Model:             ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
+		SystemFingerprint: "",
+		Object:            objectChatCompletion,
+		Choices:           []chatCompletionChoice{choice},
+		Usage: usage{
+			PromptTokens:     origResponse.Usage.InputTokens,
+			CompletionTokens: origResponse.Usage.OutputTokens,
+			TotalTokens:      origResponse.Usage.InputTokens + origResponse.Usage.OutputTokens,
+		},
+	}
+}
+
+func stopReasonClaude2OpenAI(reason *string) string {
+	if reason == nil {
+		return ""
+	}
+	switch *reason {
+	case "end_turn":
+		return finishReasonStop
+	case "stop_sequence":
+		return finishReasonStop
+	case "max_tokens":
+		return finishReasonLength
+	default:
+		return *reason
+	}
+}
+
+func (c *claudeProvider) streamResponseClaude2OpenAI(ctx wrapper.HttpContext, origResponse *claudeTextGenStreamResponse, log wrapper.Log) *chatCompletionResponse {
+	switch origResponse.Type {
+	case "message_start":
+		choice := chatCompletionChoice{
+			Index: 0,
+			Delta: &chatMessage{Role: roleAssistant, Content: ""},
+		}
+		return createChatCompletionResponse(ctx, origResponse, choice)
+
+	case "content_block_delta":
+		choice := chatCompletionChoice{
+			Index: 0,
+			Delta: &chatMessage{Content: origResponse.Delta.Text},
+		}
+		return createChatCompletionResponse(ctx, origResponse, choice)
+
+	case "message_delta":
+		choice := chatCompletionChoice{
+			Index:        0,
+			Delta:        &chatMessage{},
+			FinishReason: stopReasonClaude2OpenAI(origResponse.Delta.StopReason),
+		}
+		return createChatCompletionResponse(ctx, origResponse, choice)
+	case "content_block_stop", "message_stop":
+		log.Debugf("skip processing response type: %s", origResponse.Type)
+		return nil
+	default:
+		log.Errorf("Unexpected response type: %s", origResponse.Type)
+		return nil
+	}
+}
+
+func createChatCompletionResponse(ctx wrapper.HttpContext, response *claudeTextGenStreamResponse, choice chatCompletionChoice) *chatCompletionResponse {
+	return &chatCompletionResponse{
+		Id:      response.Message.Id,
+		Created: time.Now().UnixMilli() / 1000,
+		Model:   ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
+		Object:  objectChatCompletionChunk,
+		Choices: []chatCompletionChoice{choice},
+	}
+}
+
+func (c *claudeProvider) appendResponse(responseBuilder *strings.Builder, responseBody string) {
+	responseBuilder.WriteString(fmt.Sprintf("%s %s\n\n", streamDataItemKey, responseBody))
+}
--- a/plugins/wasm-go/extensions/ai-proxy/provider/cloudflare.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/cloudflare.go
@@ -0,0 +1,109 @@
+package provider
+
+import (
+	"errors"
+	"fmt"
+	"strings"
+
+	"github.com/alibaba/higress/plugins/wasm-go/extensions/ai-proxy/util"
+	"github.com/alibaba/higress/plugins/wasm-go/pkg/wrapper"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm/types"
+)
+
+const (
+	cloudflareDomain = "api.cloudflare.com"
+	// https://developers.cloudflare.com/workers-ai/configuration/open-ai-compatibility/
+	cloudflareChatCompletionPath = "/client/v4/accounts/{account_id}/ai/v1/chat/completions"
+)
+
+type cloudflareProviderInitializer struct {
+}
+
+func (c *cloudflareProviderInitializer) ValidateConfig(config ProviderConfig) error {
+	if config.apiTokens == nil || len(config.apiTokens) == 0 {
+		return errors.New("no apiToken found in provider config")
+	}
+	return nil
+}
+
+func (c *cloudflareProviderInitializer) CreateProvider(config ProviderConfig) (Provider, error) {
+	return &cloudflareProvider{
+		config:       config,
+		contextCache: createContextCache(&config),
+	}, nil
+}
+
+type cloudflareProvider struct {
+	config       ProviderConfig
+	contextCache *contextCache
+}
+
+func (c *cloudflareProvider) GetProviderType() string {
+	return providerTypeCloudflare
+}
+
+func (c *cloudflareProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+	_ = util.OverwriteRequestPath(strings.Replace(cloudflareChatCompletionPath, "{account_id}", c.config.cloudflareAccountId, 1))
+	_ = util.OverwriteRequestHost(cloudflareDomain)
+	_ = util.OverwriteRequestAuthorization("Bearer " + c.config.GetRandomToken())
+
+	_ = proxywasm.RemoveHttpRequestHeader("Accept-Encoding")
+	_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
+
+	return types.ActionContinue, nil
+}
+
+func (c *cloudflareProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+
+	request := &chatCompletionRequest{}
+	if err := decodeChatCompletionRequest(body, request); err != nil {
+		return types.ActionContinue, err
+	}
+	model := request.Model
+	if model == "" {
+		return types.ActionContinue, errors.New("missing model in chat completion request")
+	}
+	ctx.SetContext(ctxKeyOriginalRequestModel, model)
+	mappedModel := getMappedModel(model, c.config.modelMapping, log)
+	if mappedModel == "" {
+		return types.ActionContinue, errors.New("model becomes empty after applying the configured mapping")
+	}
+	request.Model = mappedModel
+	ctx.SetContext(ctxKeyFinalRequestModel, request.Model)
+
+	streaming := request.Stream
+	if streaming {
+		_ = proxywasm.ReplaceHttpRequestHeader("Accept", "text/event-stream")
+	}
+
+	if c.contextCache == nil {
+		if err := replaceJsonRequestBody(request, log); err != nil {
+			_ = util.SendResponse(500, "ai-proxy.cloudflare.transform_body_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+		}
+		return types.ActionContinue, nil
+	}
+	err := c.contextCache.GetContent(func(content string, err error) {
+		defer func() {
+			_ = proxywasm.ResumeHttpRequest()
+		}()
+		if err != nil {
+			log.Errorf("failed to load context file: %v", err)
+			_ = util.SendResponse(500, "ai-proxy.cloudflare.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+		}
+		insertContextMessage(request, content)
+		if err := replaceJsonRequestBody(request, log); err != nil {
+			_ = util.SendResponse(500, "ai-proxy.cloudflare.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+		}
+	}, log)
+	if err == nil {
+		return types.ActionPause, nil
+	}
+	return types.ActionContinue, err
+}
--- a/plugins/wasm-go/extensions/ai-proxy/provider/deepseek.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/deepseek.go
@@ -1,6 +1,7 @@
 package provider

 import (
+	"errors"
 	"fmt"

 	"github.com/alibaba/higress/plugins/wasm-go/extensions/ai-proxy/util"
@@ -20,6 +21,9 @@ type deepseekProviderInitializer struct {
 }

 func (m *deepseekProviderInitializer) ValidateConfig(config ProviderConfig) error {
+	if config.apiTokens == nil || len(config.apiTokens) == 0 {
+		return errors.New("no apiToken found in provider config")
+	}
 	return nil
 }

@@ -45,14 +49,8 @@ func (m *deepseekProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName Api
 	}
 	_ = util.OverwriteRequestPath(deepseekChatCompletionPath)
 	_ = util.OverwriteRequestHost(deepseekDomain)
-	_ = proxywasm.ReplaceHttpRequestHeader("Authorization", "Bearer "+m.config.GetRandomToken())
-
-	if m.contextCache == nil {
-		ctx.DontReadRequestBody()
-	} else {
-		_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
-	}
-
+	_ = util.OverwriteRequestAuthorization("Bearer " + m.config.GetRandomToken())
+	_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
 	return types.ActionContinue, nil
 }

@@ -73,11 +71,11 @@ func (m *deepseekProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiNam
 		}()
 		if err != nil {
 			log.Errorf("failed to load context file: %v", err)
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.deepseek.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
 		}
 		insertContextMessage(request, content)
 		if err := replaceJsonRequestBody(request, log); err != nil {
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.deepseek.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
 		}
 	}, log)
 	if err == nil {
--- a/plugins/wasm-go/extensions/ai-proxy/provider/gemini.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/gemini.go
@@ -0,0 +1,606 @@
+package provider
+
+import (
+	"encoding/json"
+	"errors"
+	"fmt"
+	"github.com/alibaba/higress/plugins/wasm-go/extensions/ai-proxy/util"
+	"github.com/alibaba/higress/plugins/wasm-go/pkg/wrapper"
+	"github.com/google/uuid"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm/types"
+	"strings"
+	"time"
+)
+
+// geminiProvider is the provider for google gemini/gemini flash service.
+
+const (
+	geminiApiKeyHeader = "x-goog-api-key"
+	geminiDomain       = "generativelanguage.googleapis.com"
+)
+
+type geminiProviderInitializer struct {
+}
+
+func (g *geminiProviderInitializer) ValidateConfig(config ProviderConfig) error {
+	if config.apiTokens == nil || len(config.apiTokens) == 0 {
+		return errors.New("no apiToken found in provider config")
+	}
+	return nil
+}
+
+func (g *geminiProviderInitializer) CreateProvider(config ProviderConfig) (Provider, error) {
+	return &geminiProvider{
+		config:       config,
+		contextCache: createContextCache(&config),
+	}, nil
+}
+
+type geminiProvider struct {
+	config       ProviderConfig
+	contextCache *contextCache
+}
+
+func (g *geminiProvider) GetProviderType() string {
+	return providerTypeGemini
+}
+
+func (g *geminiProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion && apiName != ApiNameEmbeddings {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+
+	_ = proxywasm.ReplaceHttpRequestHeader(geminiApiKeyHeader, g.config.GetRandomToken())
+	_ = util.OverwriteRequestHost(geminiDomain)
+
+	_ = proxywasm.RemoveHttpRequestHeader("Accept-Encoding")
+	_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
+
+	// Delay the header processing to allow changing streaming mode in OnRequestBody
+	return types.HeaderStopIteration, nil
+}
+
+func (g *geminiProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	if apiName == ApiNameChatCompletion {
+		return g.onChatCompletionRequestBody(ctx, body, log)
+	} else if apiName == ApiNameEmbeddings {
+		return g.onEmbeddingsRequestBody(ctx, body, log)
+	}
+	return types.ActionContinue, errUnsupportedApiName
+}
+
+func (g *geminiProvider) onChatCompletionRequestBody(ctx wrapper.HttpContext, body []byte, log wrapper.Log) (types.Action, error) {
+	// 使用gemini接口协议
+	if g.config.protocol == protocolOriginal {
+		request := &geminiChatRequest{}
+		if err := json.Unmarshal(body, request); err != nil {
+			return types.ActionContinue, fmt.Errorf("unable to unmarshal request: %v", err)
+		}
+		if request.Model == "" {
+			return types.ActionContinue, errors.New("request model is empty")
+		}
+		// 根据模型重写requestPath
+		path := g.getRequestPath(ApiNameChatCompletion, request.Model, request.Stream)
+		_ = util.OverwriteRequestPath(path)
+
+		// 移除多余的model和stream字段
+		request = &geminiChatRequest{
+			Contents:         request.Contents,
+			SafetySettings:   request.SafetySettings,
+			GenerationConfig: request.GenerationConfig,
+			Tools:            request.Tools,
+		}
+		if g.config.context == nil {
+			return types.ActionContinue, replaceJsonRequestBody(request, log)
+		}
+
+		err := g.contextCache.GetContent(func(content string, err error) {
+			defer func() {
+				_ = proxywasm.ResumeHttpRequest()
+			}()
+
+			if err != nil {
+				log.Errorf("failed to load context file: %v", err)
+				_ = util.SendResponse(500, "ai-proxy.gemini.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+			}
+			g.setSystemContent(request, content)
+			if err := replaceJsonRequestBody(request, log); err != nil {
+				_ = util.SendResponse(500, "ai-proxy.gemini.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+			}
+		}, log)
+		if err == nil {
+			return types.ActionPause, nil
+		}
+		return types.ActionContinue, err
+	}
+	request := &chatCompletionRequest{}
+	if err := decodeChatCompletionRequest(body, request); err != nil {
+		return types.ActionContinue, err
+	}
+
+	// 映射模型重写requestPath
+	model := request.Model
+	if model == "" {
+		return types.ActionContinue, errors.New("missing model in chat completion request")
+	}
+	ctx.SetContext(ctxKeyOriginalRequestModel, model)
+	mappedModel := getMappedModel(model, g.config.modelMapping, log)
+	if mappedModel == "" {
+		return types.ActionContinue, errors.New("model becomes empty after applying the configured mapping")
+	}
+	request.Model = mappedModel
+	ctx.SetContext(ctxKeyFinalRequestModel, request.Model)
+	path := g.getRequestPath(ApiNameChatCompletion, mappedModel, request.Stream)
+	_ = util.OverwriteRequestPath(path)
+
+	if g.config.context == nil {
+		geminiRequest := g.buildGeminiChatRequest(request)
+		return types.ActionContinue, replaceJsonRequestBody(geminiRequest, log)
+	}
+
+	err := g.contextCache.GetContent(func(content string, err error) {
+		defer func() {
+			_ = proxywasm.ResumeHttpRequest()
+		}()
+		if err != nil {
+			log.Errorf("failed to load context file: %v", err)
+			_ = util.SendResponse(500, "ai-proxy.gemini.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+		}
+		insertContextMessage(request, content)
+		geminiRequest := g.buildGeminiChatRequest(request)
+		if err := replaceJsonRequestBody(geminiRequest, log); err != nil {
+			_ = util.SendResponse(500, "ai-proxy.gemini.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+		}
+	}, log)
+	if err == nil {
+		return types.ActionPause, nil
+	}
+	return types.ActionContinue, err
+}
+
+func (g *geminiProvider) onEmbeddingsRequestBody(ctx wrapper.HttpContext, body []byte, log wrapper.Log) (types.Action, error) {
+	// 使用gemini接口协议
+	if g.config.protocol == protocolOriginal {
+		request := &geminiBatchEmbeddingRequest{}
+		if err := json.Unmarshal(body, request); err != nil {
+			return types.ActionContinue, fmt.Errorf("unable to unmarshal request: %v", err)
+		}
+		if request.Model == "" {
+			return types.ActionContinue, errors.New("request model is empty")
+		}
+		// 根据模型重写requestPath
+		path := g.getRequestPath(ApiNameEmbeddings, request.Model, false)
+		_ = util.OverwriteRequestPath(path)
+
+		// 移除多余的model字段
+		request = &geminiBatchEmbeddingRequest{
+			Requests: request.Requests,
+		}
+		return types.ActionContinue, replaceJsonRequestBody(request, log)
+	}
+	request := &embeddingsRequest{}
+	if err := json.Unmarshal(body, request); err != nil {
+		return types.ActionContinue, fmt.Errorf("unable to unmarshal request: %v", err)
+	}
+
+	// 映射模型重写requestPath
+	model := request.Model
+	if model == "" {
+		return types.ActionContinue, errors.New("missing model in embeddings request")
+	}
+	ctx.SetContext(ctxKeyOriginalRequestModel, model)
+	mappedModel := getMappedModel(model, g.config.modelMapping, log)
+	if mappedModel == "" {
+		return types.ActionContinue, errors.New("model becomes empty after applying the configured mapping")
+	}
+	request.Model = mappedModel
+	ctx.SetContext(ctxKeyFinalRequestModel, request.Model)
+	path := g.getRequestPath(ApiNameEmbeddings, mappedModel, false)
+	_ = util.OverwriteRequestPath(path)
+
+	geminiRequest := g.buildBatchEmbeddingRequest(request)
+	return types.ActionContinue, replaceJsonRequestBody(geminiRequest, log)
+}
+
+func (g *geminiProvider) OnResponseHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	if g.config.protocol == protocolOriginal {
+		ctx.DontReadResponseBody()
+		return types.ActionContinue, nil
+	}
+
+	_ = proxywasm.RemoveHttpResponseHeader("Content-Length")
+	return types.ActionContinue, nil
+}
+
+func (g *geminiProvider) OnStreamingResponseBody(ctx wrapper.HttpContext, name ApiName, chunk []byte, isLastChunk bool, log wrapper.Log) ([]byte, error) {
+	log.Infof("chunk body:%s", string(chunk))
+	if isLastChunk || len(chunk) == 0 {
+		return nil, nil
+	}
+	// sample end event response:
+	// data: {"candidates": [{"content": {"parts": [{"text": "我是 Gemini，一个大型多模态模型，由 Google 训练。我的职责是尽我所能帮助您，并尽力提供全面且信息丰富的答复。"}],"role": "model"},"finishReason": "STOP","index": 0,"safetyRatings": [{"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT","probability": "NEGLIGIBLE"},{"category": "HARM_CATEGORY_HATE_SPEECH","probability": "NEGLIGIBLE"},{"category": "HARM_CATEGORY_HARASSMENT","probability": "NEGLIGIBLE"},{"category": "HARM_CATEGORY_DANGEROUS_CONTENT","probability": "NEGLIGIBLE"}]}],"usageMetadata": {"promptTokenCount": 2,"candidatesTokenCount": 35,"totalTokenCount": 37}}
+	responseBuilder := &strings.Builder{}
+	lines := strings.Split(string(chunk), "\n")
+	for _, data := range lines {
+		if len(data) < 6 {
+			// ignore blank line or wrong format
+			continue
+		}
+		data = data[6:]
+		var geminiResp geminiChatResponse
+		if err := json.Unmarshal([]byte(data), &geminiResp); err != nil {
+			log.Errorf("unable to unmarshal gemini response: %v", err)
+			continue
+		}
+		response := g.buildChatCompletionStreamResponse(ctx, &geminiResp)
+		responseBody, err := json.Marshal(response)
+		if err != nil {
+			log.Errorf("unable to marshal response: %v", err)
+			return nil, err
+		}
+		g.appendResponse(responseBuilder, string(responseBody))
+	}
+	modifiedResponseChunk := responseBuilder.String()
+	log.Debugf("=== modified response chunk: %s", modifiedResponseChunk)
+	return []byte(modifiedResponseChunk), nil
+}
+
+func (g *geminiProvider) OnResponseBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	if apiName == ApiNameChatCompletion {
+		return g.onChatCompletionResponseBody(ctx, body, log)
+	} else if apiName == ApiNameEmbeddings {
+		return g.onEmbeddingsResponseBody(ctx, body, log)
+	}
+	return types.ActionContinue, errUnsupportedApiName
+}
+
+func (g *geminiProvider) onChatCompletionResponseBody(ctx wrapper.HttpContext, body []byte, log wrapper.Log) (types.Action, error) {
+	geminiResponse := &geminiChatResponse{}
+	if err := json.Unmarshal(body, geminiResponse); err != nil {
+		return types.ActionContinue, fmt.Errorf("unable to unmarshal gemini chat response: %v", err)
+	}
+	if geminiResponse.Error != nil {
+		return types.ActionContinue, fmt.Errorf("gemini chat completion response error, error_code: %d, error_status:%s, error_message: %s",
+			geminiResponse.Error.Code, geminiResponse.Error.Status, geminiResponse.Error.Message)
+	}
+	response := g.buildChatCompletionResponse(ctx, geminiResponse)
+	return types.ActionContinue, replaceJsonResponseBody(response, log)
+}
+
+func (g *geminiProvider) onEmbeddingsResponseBody(ctx wrapper.HttpContext, body []byte, log wrapper.Log) (types.Action, error) {
+	geminiResponse := &geminiEmbeddingResponse{}
+	if err := json.Unmarshal(body, geminiResponse); err != nil {
+		return types.ActionContinue, fmt.Errorf("unable to unmarshal gemini embeddings response: %v", err)
+	}
+	if geminiResponse.Error != nil {
+		return types.ActionContinue, fmt.Errorf("gemini embeddings response error, error_code: %d, error_status:%s, error_message: %s",
+			geminiResponse.Error.Code, geminiResponse.Error.Status, geminiResponse.Error.Message)
+	}
+	response := g.buildEmbeddingsResponse(ctx, geminiResponse)
+	return types.ActionContinue, replaceJsonResponseBody(response, log)
+}
+
+func (g *geminiProvider) getRequestPath(apiName ApiName, geminiModel string, stream bool) string {
+	action := ""
+	if apiName == ApiNameEmbeddings {
+		action = "batchEmbedContents"
+	} else if stream {
+		action = "streamGenerateContent?alt=sse"
+	} else {
+		action = "generateContent"
+	}
+	return fmt.Sprintf("/v1/models/%s:%s", geminiModel, action)
+}
+
+type geminiChatRequest struct {
+	// Model and Stream are only used when using the gemini original protocol
+	Model            string                     `json:"model,omitempty"`
+	Stream           bool                       `json:"stream,omitempty"`
+	Contents         []geminiChatContent        `json:"contents"`
+	SafetySettings   []geminiChatSafetySetting  `json:"safety_settings,omitempty"`
+	GenerationConfig geminiChatGenerationConfig `json:"generation_config,omitempty"`
+	Tools            []geminiChatTools          `json:"tools,omitempty"`
+}
+
+type geminiChatContent struct {
+	Role  string       `json:"role,omitempty"`
+	Parts []geminiPart `json:"parts"`
+}
+
+type geminiChatSafetySetting struct {
+	Category  string `json:"category"`
+	Threshold string `json:"threshold"`
+}
+
+type geminiChatGenerationConfig struct {
+	Temperature     float64  `json:"temperature,omitempty"`
+	TopP            float64  `json:"topP,omitempty"`
+	TopK            float64  `json:"topK,omitempty"`
+	MaxOutputTokens int      `json:"maxOutputTokens,omitempty"`
+	CandidateCount  int      `json:"candidateCount,omitempty"`
+	StopSequences   []string `json:"stopSequences,omitempty"`
+}
+
+type geminiChatTools struct {
+	FunctionDeclarations any `json:"function_declarations,omitempty"`
+}
+
+type geminiPart struct {
+	Text         string              `json:"text,omitempty"`
+	InlineData   *geminiInlineData   `json:"inlineData,omitempty"`
+	FunctionCall *geminiFunctionCall `json:"functionCall,omitempty"`
+}
+
+type geminiInlineData struct {
+	MimeType string `json:"mimeType"`
+	Data     string `json:"data"`
+}
+
+type geminiFunctionCall struct {
+	FunctionName string `json:"name"`
+	Arguments    any    `json:"args"`
+}
+
+func (g *geminiProvider) buildGeminiChatRequest(request *chatCompletionRequest) *geminiChatRequest {
+	var safetySettings []geminiChatSafetySetting
+	{
+	}
+	for category, threshold := range g.config.geminiSafetySetting {
+		safetySettings = append(safetySettings, geminiChatSafetySetting{
+			Category:  category,
+			Threshold: threshold,
+		})
+	}
+	geminiRequest := geminiChatRequest{
+		Contents:       make([]geminiChatContent, 0, len(request.Messages)),
+		SafetySettings: safetySettings,
+		GenerationConfig: geminiChatGenerationConfig{
+			Temperature:     request.Temperature,
+			TopP:            request.TopP,
+			MaxOutputTokens: request.MaxTokens,
+		},
+	}
+	if request.Tools != nil {
+		functions := make([]function, 0, len(request.Tools))
+		for _, tool := range request.Tools {
+			functions = append(functions, tool.Function)
+		}
+		geminiRequest.Tools = []geminiChatTools{
+			{
+				FunctionDeclarations: functions,
+			},
+		}
+	}
+	shouldAddDummyModelMessage := false
+	for _, message := range request.Messages {
+		content := geminiChatContent{
+			Role: message.Role,
+			Parts: []geminiPart{
+				{
+					Text: message.Content,
+				},
+			},
+		}
+
+		// there's no assistant role in gemini and API shall vomit if role is not user or model
+		if content.Role == roleAssistant {
+			content.Role = "model"
+		} else if content.Role == roleSystem { // converting system prompt to prompt from user for the same reason
+			content.Role = roleUser
+			shouldAddDummyModelMessage = true
+		}
+		geminiRequest.Contents = append(geminiRequest.Contents, content)
+
+		// if a system message is the last message, we need to add a dummy model message to make gemini happy
+		if shouldAddDummyModelMessage {
+			geminiRequest.Contents = append(geminiRequest.Contents, geminiChatContent{
+				Role: "model",
+				Parts: []geminiPart{
+					{
+						Text: "Okay",
+					},
+				},
+			})
+			shouldAddDummyModelMessage = false
+		}
+	}
+
+	return &geminiRequest
+}
+
+func (g *geminiProvider) setSystemContent(request *geminiChatRequest, content string) {
+	systemContents := []geminiChatContent{{
+		Role: roleUser,
+		Parts: []geminiPart{
+			{
+				Text: content,
+			},
+		},
+	}}
+	request.Contents = append(systemContents, request.Contents...)
+}
+
+type geminiBatchEmbeddingRequest struct {
+	// Model are only used when using the gemini original protocol
+	Model    string                   `json:"model,omitempty"`
+	Requests []geminiEmbeddingRequest `json:"requests"`
+}
+
+type geminiEmbeddingRequest struct {
+	Model                string            `json:"model"`
+	Content              geminiChatContent `json:"content"`
+	TaskType             string            `json:"taskType,omitempty"`
+	Title                string            `json:"title,omitempty"`
+	OutputDimensionality int               `json:"outputDimensionality,omitempty"`
+}
+
+func (g *geminiProvider) buildBatchEmbeddingRequest(request *embeddingsRequest) *geminiBatchEmbeddingRequest {
+	inputs := request.ParseInput()
+	requests := make([]geminiEmbeddingRequest, len(inputs))
+	model := fmt.Sprintf("models/%s", request.Model)
+
+	for i, input := range inputs {
+		requests[i] = geminiEmbeddingRequest{
+			Model: model,
+			Content: geminiChatContent{
+				Parts: []geminiPart{
+					{
+						Text: input,
+					},
+				},
+			},
+		}
+	}
+
+	return &geminiBatchEmbeddingRequest{
+		Requests: requests,
+	}
+}
+
+type geminiChatResponse struct {
+	Candidates     []geminiChatCandidate    `json:"candidates"`
+	PromptFeedback geminiChatPromptFeedback `json:"promptFeedback"`
+	UsageMetadata  geminiUsageMetadata      `json:"usageMetadata"`
+	Error          *geminiResponseError     `json:"error,omitempty"`
+}
+
+type geminiChatCandidate struct {
+	Content       geminiChatContent        `json:"content"`
+	FinishReason  string                   `json:"finishReason"`
+	Index         int64                    `json:"index"`
+	SafetyRatings []geminiChatSafetyRating `json:"safetyRatings"`
+}
+
+type geminiChatPromptFeedback struct {
+	SafetyRatings []geminiChatSafetyRating `json:"safetyRatings"`
+}
+
+type geminiUsageMetadata struct {
+	PromptTokenCount     int `json:"promptTokenCount,omitempty"`
+	CandidatesTokenCount int `json:"candidatesTokenCount,omitempty"`
+	TotalTokenCount      int `json:"totalTokenCount,omitempty"`
+}
+
+type geminiResponseError struct {
+	Code    int    `json:"code,omitempty"`
+	Message string `json:"message,omitempty"`
+	Status  string `json:"status,omitempty"`
+}
+
+type geminiChatSafetyRating struct {
+	Category    string `json:"category"`
+	Probability string `json:"probability"`
+}
+
+func (g *geminiProvider) buildChatCompletionResponse(ctx wrapper.HttpContext, response *geminiChatResponse) *chatCompletionResponse {
+	fullTextResponse := chatCompletionResponse{
+		Id:      fmt.Sprintf("chatcmpl-%s", uuid.New().String()),
+		Object:  objectChatCompletion,
+		Created: time.Now().UnixMilli() / 1000,
+		Model:   ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
+		Choices: make([]chatCompletionChoice, 0, len(response.Candidates)),
+		Usage: usage{
+			PromptTokens:     response.UsageMetadata.PromptTokenCount,
+			CompletionTokens: response.UsageMetadata.CandidatesTokenCount,
+			TotalTokens:      response.UsageMetadata.TotalTokenCount,
+		},
+	}
+	for i, candidate := range response.Candidates {
+		choice := chatCompletionChoice{
+			Index: i,
+			Message: &chatMessage{
+				Role: roleAssistant,
+			},
+			FinishReason: finishReasonStop,
+		}
+		if len(candidate.Content.Parts) > 0 {
+			if candidate.Content.Parts[0].FunctionCall != nil {
+				choice.Message.ToolCalls = g.buildToolCalls(&candidate)
+			} else {
+				choice.Message.Content = candidate.Content.Parts[0].Text
+			}
+		} else {
+			choice.Message.Content = ""
+			choice.FinishReason = candidate.FinishReason
+		}
+		fullTextResponse.Choices = append(fullTextResponse.Choices, choice)
+	}
+	return &fullTextResponse
+}
+
+func (g *geminiProvider) buildToolCalls(candidate *geminiChatCandidate) []toolCall {
+	var toolCalls []toolCall
+
+	item := candidate.Content.Parts[0]
+	if item.FunctionCall != nil {
+		return toolCalls
+	}
+	argsBytes, err := json.Marshal(item.FunctionCall.Arguments)
+	if err != nil {
+		proxywasm.LogErrorf("get toolCalls from gemini response failed: " + err.Error())
+		return toolCalls
+	}
+	toolCall := toolCall{
+		Id:   fmt.Sprintf("call_%s", uuid.New().String()),
+		Type: "function",
+		Function: functionCall{
+			Arguments: string(argsBytes),
+			Name:      item.FunctionCall.FunctionName,
+		},
+	}
+	toolCalls = append(toolCalls, toolCall)
+	return toolCalls
+}
+
+func (g *geminiProvider) buildChatCompletionStreamResponse(ctx wrapper.HttpContext, geminiResp *geminiChatResponse) *chatCompletionResponse {
+	var choice chatCompletionChoice
+	if len(geminiResp.Candidates) > 0 && len(geminiResp.Candidates[0].Content.Parts) > 0 {
+		choice.Delta = &chatMessage{Content: geminiResp.Candidates[0].Content.Parts[0].Text}
+	}
+	streamResponse := chatCompletionResponse{
+		Id:      fmt.Sprintf("chatcmpl-%s", uuid.New().String()),
+		Object:  objectChatCompletionChunk,
+		Created: time.Now().UnixMilli() / 1000,
+		Model:   ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
+		Choices: []chatCompletionChoice{choice},
+		Usage: usage{
+			PromptTokens:     geminiResp.UsageMetadata.PromptTokenCount,
+			CompletionTokens: geminiResp.UsageMetadata.CandidatesTokenCount,
+			TotalTokens:      geminiResp.UsageMetadata.TotalTokenCount,
+		},
+	}
+	return &streamResponse
+}
+
+type geminiEmbeddingResponse struct {
+	Embeddings []geminiEmbeddingData `json:"embeddings"`
+	Error      *geminiResponseError  `json:"error,omitempty"`
+}
+
+type geminiEmbeddingData struct {
+	Values []float64 `json:"values"`
+}
+
+func (g *geminiProvider) buildEmbeddingsResponse(ctx wrapper.HttpContext, geminiResp *geminiEmbeddingResponse) *embeddingsResponse {
+	response := embeddingsResponse{
+		Object: "list",
+		Data:   make([]embedding, 0, len(geminiResp.Embeddings)),
+		Model:  ctx.GetContext(ctxKeyFinalRequestModel).(string),
+		Usage: usage{
+			TotalTokens: 0,
+		},
+	}
+	for _, item := range geminiResp.Embeddings {
+		response.Data = append(response.Data, embedding{
+			Object:    `embedding`,
+			Index:     0,
+			Embedding: item.Values,
+		})
+	}
+	return &response
+}
+
+func (g *geminiProvider) appendResponse(responseBuilder *strings.Builder, responseBody string) {
+	responseBuilder.WriteString(fmt.Sprintf("%s %s\n\n", streamDataItemKey, responseBody))
+}
--- a/plugins/wasm-go/extensions/ai-proxy/provider/groq.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/groq.go
@@ -1,6 +1,7 @@
 package provider

 import (
+	"errors"
 	"fmt"

 	"github.com/alibaba/higress/plugins/wasm-go/extensions/ai-proxy/util"
@@ -18,6 +19,9 @@ const (
 type groqProviderInitializer struct{}

 func (m *groqProviderInitializer) ValidateConfig(config ProviderConfig) error {
+	if config.apiTokens == nil || len(config.apiTokens) == 0 {
+		return errors.New("no apiToken found in provider config")
+	}
 	return nil
 }

@@ -43,14 +47,8 @@ func (m *groqProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName ApiName
 	}
 	_ = util.OverwriteRequestPath(groqChatCompletionPath)
 	_ = util.OverwriteRequestHost(groqDomain)
-	_ = proxywasm.ReplaceHttpRequestHeader("Authorization", "Bearer "+m.config.GetRandomToken())
-
-	if m.contextCache == nil {
-		ctx.DontReadRequestBody()
-	} else {
-		_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
-	}
-
+	_ = util.OverwriteRequestAuthorization("Bearer " + m.config.GetRandomToken())
+	_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
 	return types.ActionContinue, nil
 }

@@ -71,11 +69,11 @@ func (m *groqProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, b
 		}()
 		if err != nil {
 			log.Errorf("failed to load context file: %v", err)
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.groq.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
 		}
 		insertContextMessage(request, content)
 		if err := replaceJsonRequestBody(request, log); err != nil {
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.groq.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
 		}
 	}, log)
 	if err == nil {
--- a/plugins/wasm-go/extensions/ai-proxy/provider/hunyuan.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/hunyuan.go
@@ -0,0 +1,563 @@
+package provider
+
+import (
+	"bytes"
+	"crypto/hmac"
+	"crypto/sha256"
+	"encoding/hex"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"strings"
+	"time"
+
+	"github.com/alibaba/higress/plugins/wasm-go/extensions/ai-proxy/util"
+	"github.com/alibaba/higress/plugins/wasm-go/pkg/wrapper"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm/types"
+)
+
+// hunyuanProvider is the provider for hunyuan AI service.
+
+const (
+	hunyuanDomain                 = "hunyuan.tencentcloudapi.com"
+	hunyuanRequestPath            = "/"
+	hunyuanChatCompletionTCAction = "ChatCompletions"
+
+	// headers necessary for TC hunyuan api call:
+	// ref: https://cloud.tencent.com/document/api/1729/105701, https://cloud.tencent.com/document/api/1729/101842
+	actionKey        = "X-TC-Action"
+	timestampKey     = "X-TC-Timestamp"
+	authorizationKey = "Authorization"
+	versionKey       = "X-TC-Version"
+	versionValue     = "2023-09-01"
+	hostKey          = "Host"
+
+	ssePrefix            = "data: " // Server-Sent Events (SSE) 类型的流式响应的开始标记
+	hunyuanStreamEndMark = "stop"   // 混元的流式的finishReason为stop时，表示结束
+
+	hunyuanAuthKeyLen = 32
+	hunyuanAuthIdLen  = 36
+)
+
+type hunyuanProviderInitializer struct {
+}
+
+// ref: https://console.cloud.tencent.com/api/explorer?Product=hunyuan&Version=2023-09-01&Action=ChatCompletions
+type hunyuanTextGenRequest struct {
+	Model             string               `json:"Model"`
+	Messages          []hunyuanChatMessage `json:"Messages"`
+	Stream            bool                 `json:"Stream,omitempty"`
+	StreamModeration  bool                 `json:"StreamModeration,omitempty"`
+	TopP              float32              `json:"TopP,omitempty"`
+	Temperature       float32              `json:"Temperature,omitempty"`
+	EnableEnhancement bool                 `json:"EnableEnhancement,omitempty"`
+}
+
+type hunyuanTextGenResponseNonStreaming struct {
+	Response hunyuanTextGenDetailedResponseNonStreaming `json:"Response"`
+}
+
+type hunyuanTextGenDetailedResponseNonStreaming struct {
+	RequestId string                 `json:"RequestId,omitempty"`
+	Note      string                 `json:"Note"`
+	Choices   []hunyuanTextGenChoice `json:"Choices"`
+	Created   int64                  `json:"Created"`
+	Id        string                 `json:"Id"`
+	Usage     hunyuanTextGenUsage    `json:"Usage"`
+}
+
+type hunyuanTextGenChoice struct {
+	FinishReason string             `json:"FinishReason"`
+	Message      hunyuanChatMessage `json:"Message,omitempty"` // 当非流式返回时存储大模型生成文字
+	Delta        hunyuanChatMessage `json:"Delta,omitempty"`   // 流式返回时存储大模型生成文字
+}
+
+type hunyuanTextGenUsage struct {
+	PromptTokens     int `json:"PromptTokens"`
+	CompletionTokens int `json:"CompletionTokens"`
+	TotalTokens      int `json:"TotalTokens"`
+}
+
+type hunyuanChatMessage struct {
+	Role    string `json:"Role,omitempty"`
+	Content string `json:"Content,omitempty"`
+}
+
+func (m *hunyuanProviderInitializer) ValidateConfig(config ProviderConfig) error {
+	// 校验hunyuan id 和 key的合法性
+	if len(config.hunyuanAuthId) != hunyuanAuthIdLen || len(config.hunyuanAuthKey) != hunyuanAuthKeyLen {
+		return errors.New("hunyuanAuthId / hunyuanAuthKey is illegal in config file")
+	}
+	return nil
+}
+
+func (m *hunyuanProviderInitializer) CreateProvider(config ProviderConfig) (Provider, error) {
+	return &hunyuanProvider{
+		config: config,
+		client: wrapper.NewClusterClient(wrapper.RouteCluster{
+			Host: hunyuanDomain,
+		}),
+		contextCache: createContextCache(&config),
+	}, nil
+}
+
+type hunyuanProvider struct {
+	config ProviderConfig
+
+	client       wrapper.HttpClient
+	contextCache *contextCache
+}
+
+func (m *hunyuanProvider) GetProviderType() string {
+	return providerTypeHunyuan
+}
+
+func (m *hunyuanProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	// log.Debugf("hunyuanProvider.OnRequestHeaders called! hunyunSecretKey/id is: %s/%s", m.config.hunyuanAuthKey, m.config.hunyuanAuthId)
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+
+	_ = util.OverwriteRequestHost(hunyuanDomain)
+	_ = util.OverwriteRequestPath(hunyuanRequestPath)
+
+	// 添加hunyuan需要的自定义字段
+	_ = proxywasm.ReplaceHttpRequestHeader(actionKey, hunyuanChatCompletionTCAction)
+	_ = proxywasm.ReplaceHttpRequestHeader(versionKey, versionValue)
+
+	// 删除一些字段
+	_ = proxywasm.RemoveHttpRequestHeader("Accept-Encoding")
+	_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
+
+	// Delay the header processing to allow changing streaming mode in OnRequestBody
+	return types.HeaderStopIteration, nil
+}
+
+func (m *hunyuanProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+
+	// 为header添加时间戳字段 （因为需要根据body进行签名时依赖时间戳，故于body处理部分创建时间戳）
+	var timestamp int64 = time.Now().Unix()
+	_ = proxywasm.ReplaceHttpRequestHeader(timestampKey, fmt.Sprintf("%d", timestamp))
+	// log.Debugf("#debug nash5# OnRequestBody set timestamp header: ", timestamp)
+
+	// 使用混元本身接口的协议
+	if m.config.protocol == protocolOriginal {
+		request := &hunyuanTextGenRequest{}
+
+		if err := json.Unmarshal(body, request); err != nil {
+			return types.ActionContinue, fmt.Errorf("unable to unmarshal request: %v", err)
+		}
+
+		// 根据确定好的payload进行签名
+		hunyuanBody, _ := json.Marshal(request)
+		authorizedValueNew := GetTC3Authorizationcode(m.config.hunyuanAuthId, m.config.hunyuanAuthKey, timestamp, hunyuanDomain, hunyuanChatCompletionTCAction, string(hunyuanBody))
+		_ = util.OverwriteRequestAuthorization(authorizedValueNew)
+		_ = proxywasm.ReplaceHttpRequestHeader("Accept", "*/*")
+		// log.Debugf("#debug nash5# OnRequestBody call hunyuan api using original api! signature computation done!")
+
+		// 若无配置文件，直接返回
+		if m.config.context == nil {
+			return types.ActionContinue, replaceJsonRequestBody(request, log)
+		}
+		err := m.contextCache.GetContent(func(content string, err error) {
+			log.Debugf("#debug nash5# ctx file loaded! callback start, content is: %s", content)
+			defer func() {
+				_ = proxywasm.ResumeHttpRequest()
+			}()
+
+			if err != nil {
+				log.Errorf("failed to load context file: %v", err)
+				_ = util.SendResponse(500, "ai-proxy.hunyuan.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+			}
+			m.insertContextMessageIntoHunyuanRequest(request, content)
+
+			// 因为手动插入了context内容，这里需要重新计算签名
+			hunyuanBody, _ := json.Marshal(request)
+			authorizedValueNew := GetTC3Authorizationcode(m.config.hunyuanAuthId, m.config.hunyuanAuthKey, timestamp, hunyuanDomain, hunyuanChatCompletionTCAction, string(hunyuanBody))
+			_ = util.OverwriteRequestAuthorization(authorizedValueNew)
+
+			if err := replaceJsonRequestBody(request, log); err != nil {
+				_ = util.SendResponse(500, "ai-proxy.hunyuan.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+			}
+		}, log)
+		if err == nil {
+			log.Debugf("#debug nash5# ctx file load success!")
+			return types.ActionPause, nil
+		}
+
+		log.Debugf("#debug nash5# ctx file load failed!")
+		return types.ActionContinue, replaceJsonRequestBody(request, log)
+	}
+
+	// 使用open ai接口协议
+	request := &chatCompletionRequest{}
+	if err := decodeChatCompletionRequest(body, request); err != nil {
+		return types.ActionContinue, err
+	}
+	// log.Debugf("#debug nash5# OnRequestBody call hunyuan api using openai's api!")
+
+	model := request.Model
+	if model == "" {
+		return types.ActionContinue, errors.New("missing model in chat completion request")
+	}
+	ctx.SetContext(ctxKeyOriginalRequestModel, model) // 设置原始请求的model，以便返回值使用
+	mappedModel := getMappedModel(model, m.config.modelMapping, log)
+	if mappedModel == "" {
+		return types.ActionContinue, errors.New("model becomes empty after applying the configured mapping")
+	}
+	request.Model = mappedModel
+	ctx.SetContext(ctxKeyFinalRequestModel, request.Model) // 设置真实请求的模型，以便返回值使用
+
+	// 看请求中的stream的设置，相应的我们更该http头
+	streaming := request.Stream
+	if streaming {
+		_ = proxywasm.ReplaceHttpRequestHeader("Accept", "text/event-stream")
+	} else {
+		_ = proxywasm.ReplaceHttpRequestHeader("Accept", "*/*")
+	}
+
+	// 若没有配置上下文，直接开始请求
+	if m.config.context == nil {
+		hunyuanRequest := m.buildHunyuanTextGenerationRequest(request)
+
+		// 根据确定好的payload进行签名：
+		body, _ := json.Marshal(hunyuanRequest)
+		authorizedValueNew := GetTC3Authorizationcode(
+			m.config.hunyuanAuthId,
+			m.config.hunyuanAuthKey,
+			timestamp,
+			hunyuanDomain,
+			hunyuanChatCompletionTCAction,
+			string(body),
+		)
+		_ = util.OverwriteRequestAuthorization(authorizedValueNew)
+		// log.Debugf("#debug nash5# OnRequestBody done, body is: ", string(body))
+
+		// // 打印所有的headers
+		// headers, err2 := proxywasm.GetHttpRequestHeaders()
+		// if err2 != nil {
+		// 	log.Errorf("failed to get request headers: %v", err2)
+		// } else {
+		// 	// 迭代并打印所有请求头
+		// 	for _, header := range headers {
+		// 		log.Infof("#debug nash5# inB Request header - %s: %s", header[0], header[1])
+		// 	}
+		// }
+		return types.ActionContinue, replaceJsonRequestBody(hunyuanRequest, log)
+	}
+
+	err := m.contextCache.GetContent(func(content string, err error) {
+		defer func() {
+			_ = proxywasm.ResumeHttpRequest()
+		}()
+		if err != nil {
+			log.Errorf("failed to load context file: %v", err)
+			_ = util.SendResponse(500, "ai-proxy.hunyuan.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+			return
+		}
+		insertContextMessage(request, content)
+		hunyuanRequest := m.buildHunyuanTextGenerationRequest(request)
+
+		// 因为手动插入了context内容，这里需要重新计算签名
+		hunyuanBody, _ := json.Marshal(hunyuanRequest)
+		authorizedValueNew := GetTC3Authorizationcode(m.config.hunyuanAuthId, m.config.hunyuanAuthKey, timestamp, hunyuanDomain, hunyuanChatCompletionTCAction, string(hunyuanBody))
+		_ = util.OverwriteRequestAuthorization(authorizedValueNew)
+
+		if err := replaceJsonRequestBody(hunyuanRequest, log); err != nil {
+			_ = util.SendResponse(500, "ai-proxy.hunyuan.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+		}
+	}, log)
+	if err == nil {
+		return types.ActionPause, nil
+	}
+	return types.ActionContinue, err
+}
+
+func (m *hunyuanProvider) OnResponseHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	_ = proxywasm.RemoveHttpResponseHeader("Content-Length")
+	return types.ActionContinue, nil
+}
+
+func (m *hunyuanProvider) OnStreamingResponseBody(ctx wrapper.HttpContext, name ApiName, chunk []byte, isLastChunk bool, log wrapper.Log) ([]byte, error) {
+	if m.config.protocol == protocolOriginal {
+		return chunk, nil
+	}
+
+	// hunyuan的流式返回:
+	//data: {"Note":"以上内容为AI生成，不代表开发者立场，请勿删除或修改本标记","Choices":[{"Delta":{"Role":"assistant","Content":"有助于"},"FinishReason":""}],"Created":1716359713,"Id":"086b6b19-8b2c-4def-a65c-db6a7bc86acd","Usage":{"PromptTokens":7,"CompletionTokens":145,"TotalTokens":152}}
+
+	// openai的流式返回
+	// data: {"id": "chatcmpl-7QyqpwdfhqwajicIEznoc6Q47XAyW", "object": "chat.completion.chunk", "created": 1677664795, "model": "gpt-3.5-turbo-0613", "choices": [{"delta": {"content": "The "}, "index": 0, "finish_reason": null}]}
+
+	// log.Debugf("#debug nash5# [OnStreamingResponseBody] chunk is: %s", string(chunk))
+
+	// 从上下文获取现有缓冲区数据
+	newBufferedBody := chunk
+	if bufferedBody, has := ctx.GetContext(ctxKeyStreamingBody).([]byte); has {
+		newBufferedBody = append(bufferedBody, chunk...)
+	}
+
+	// 初始化处理下标，以及将要返回的处理过的chunks
+	var newEventPivot = -1
+	var outputBuffer []byte
+
+	// 从buffer区取出若干完整的chunk，将其转为openAI格式后返回
+	// 处理可能包含多个事件的缓冲区
+	for {
+		eventStartIndex := bytes.Index(newBufferedBody, []byte(ssePrefix))
+		if eventStartIndex == -1 {
+			break // 没有找到新事件，跳出循环
+		}
+
+		// 移除缓冲区前面非事件部分
+		newBufferedBody = newBufferedBody[eventStartIndex+len(ssePrefix):]
+
+		// 查找事件结束的位置（即下一个事件的开始）
+		newEventPivot = bytes.Index(newBufferedBody, []byte("\n\n"))
+		if newEventPivot == -1 && !isLastChunk {
+			// 未找到事件结束标识，跳出循环等待更多数据，若是最后一个chunk，不一定有2个换行符
+			break
+		}
+
+		// 提取并处理一个完整的事件
+		eventData := newBufferedBody[:newEventPivot]
+		// log.Debugf("@@@ <<< ori chun is: %s", string(newBufferedBody[:newEventPivot]))
+		newBufferedBody = newBufferedBody[newEventPivot+2:] // 跳过结束标识
+
+		// 转换并追加到输出缓冲区
+		convertedData, _ := m.convertChunkFromHunyuanToOpenAI(ctx, eventData, log)
+		// log.Debugf("@@@ >>> converted one chunk: %s", string(convertedData))
+		outputBuffer = append(outputBuffer, convertedData...)
+	}
+
+	// 刷新剩余的不完整事件回到上下文缓冲区以便下次继续处理
+	ctx.SetContext(ctxKeyStreamingBody, newBufferedBody)
+
+	log.Debugf("=== modified response chunk: %s", string(outputBuffer))
+	return outputBuffer, nil
+}
+
+func (m *hunyuanProvider) convertChunkFromHunyuanToOpenAI(ctx wrapper.HttpContext, hunyuanChunk []byte, log wrapper.Log) ([]byte, error) {
+	// 将hunyuan的chunk转为openai的chunk
+	hunyuanFormattedChunk := &hunyuanTextGenDetailedResponseNonStreaming{}
+	if err := json.Unmarshal(hunyuanChunk, hunyuanFormattedChunk); err != nil {
+		return []byte(""), nil
+	}
+
+	openAIFormattedChunk := &chatCompletionResponse{
+		Id:                hunyuanFormattedChunk.Id,
+		Created:           time.Now().UnixMilli() / 1000,
+		Model:             ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
+		SystemFingerprint: "",
+		Object:            objectChatCompletionChunk,
+		Usage: usage{
+			PromptTokens:     hunyuanFormattedChunk.Usage.PromptTokens,
+			CompletionTokens: hunyuanFormattedChunk.Usage.CompletionTokens,
+			TotalTokens:      hunyuanFormattedChunk.Usage.TotalTokens,
+		},
+	}
+	// tmpStr3, _ := json.Marshal(hunyuanFormattedChunk)
+	// log.Debugf("@@@ --- 源数据是：: %s", tmpStr3)
+
+	// 是否为最后一个chunk？
+	if hunyuanFormattedChunk.Choices[0].FinishReason == hunyuanStreamEndMark {
+		// log.Debugf("@@@ --- 最后chunk: ")
+		openAIFormattedChunk.Choices = append(openAIFormattedChunk.Choices, chatCompletionChoice{
+			FinishReason: hunyuanFormattedChunk.Choices[0].FinishReason,
+		})
+	} else {
+		deltaMsg := chatMessage{
+			Name:      "",
+			Role:      hunyuanFormattedChunk.Choices[0].Delta.Role,
+			Content:   hunyuanFormattedChunk.Choices[0].Delta.Content,
+			ToolCalls: []toolCall{},
+		}
+
+		// tmpStr2, _ := json.Marshal(deltaMsg)
+		// log.Debugf("@@@ --- 中间chunk: choices.chatMsg 是: %s", tmpStr2)
+
+		openAIFormattedChunk.Choices = append(
+			openAIFormattedChunk.Choices,
+			chatCompletionChoice{Delta: &deltaMsg},
+		)
+		// tmpStr, _ := json.Marshal(openAIFormattedChunk.Choices)
+		// log.Debugf("@@@ --- 中间chunk: choices 是: %s", tmpStr)
+	}
+
+	// 返回的格式
+	openAIFormattedChunkBytes, _ := json.Marshal(openAIFormattedChunk)
+	var openAIChunk strings.Builder
+	openAIChunk.WriteString(ssePrefix)
+	openAIChunk.WriteString(string(openAIFormattedChunkBytes))
+	openAIChunk.WriteString("\n\n")
+
+	return []byte(openAIChunk.String()), nil
+}
+
+func (m *hunyuanProvider) OnResponseBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+
+	log.Debugf("#debug nash5# onRespBody's resp is: %s", string(body))
+	hunyuanResponse := &hunyuanTextGenResponseNonStreaming{}
+	if err := json.Unmarshal(body, hunyuanResponse); err != nil {
+		return types.ActionContinue, fmt.Errorf("unable to unmarshal hunyuan response: %v", err)
+	}
+
+	if m.config.protocol == protocolOriginal {
+		return types.ActionContinue, replaceJsonResponseBody(hunyuanResponse, log)
+	}
+
+	response := m.buildChatCompletionResponse(ctx, hunyuanResponse)
+
+	return types.ActionContinue, replaceJsonResponseBody(response, log)
+}
+
+func (m *hunyuanProvider) insertContextMessageIntoHunyuanRequest(request *hunyuanTextGenRequest, content string) {
+
+	fileMessage := hunyuanChatMessage{
+		Role:    roleSystem,
+		Content: content,
+	}
+	messages := request.Messages
+	request.Messages = append([]hunyuanChatMessage{},
+		append([]hunyuanChatMessage{fileMessage}, messages...)...,
+	)
+}
+
+func (m *hunyuanProvider) buildHunyuanTextGenerationRequest(request *chatCompletionRequest) *hunyuanTextGenRequest {
+	hunyuanRequest := &hunyuanTextGenRequest{
+		Model:             request.Model,
+		Messages:          convertMessagesFromOpenAIToHunyuan(request.Messages),
+		Stream:            request.Stream,
+		StreamModeration:  false,
+		TopP:              float32(request.TopP),
+		Temperature:       float32(request.Temperature),
+		EnableEnhancement: false,
+	}
+
+	return hunyuanRequest
+}
+
+func convertMessagesFromOpenAIToHunyuan(openAIMessages []chatMessage) []hunyuanChatMessage {
+	// 将chatgpt的messages转换为hunyuan的messages
+	hunyuanChatMessages := make([]hunyuanChatMessage, 0, len(openAIMessages))
+	for _, msg := range openAIMessages {
+		hunyuanChatMessages = append(hunyuanChatMessages, hunyuanChatMessage{
+			Role:    msg.Role,
+			Content: msg.Content,
+		})
+	}
+
+	return hunyuanChatMessages
+}
+
+func (m *hunyuanProvider) buildChatCompletionResponse(ctx wrapper.HttpContext, hunyuanResponse *hunyuanTextGenResponseNonStreaming) *chatCompletionResponse {
+	choices := make([]chatCompletionChoice, 0, len(hunyuanResponse.Response.Choices))
+	for _, choice := range hunyuanResponse.Response.Choices {
+		choices = append(choices, chatCompletionChoice{
+			Message: &chatMessage{
+				Name:      "",
+				Role:      choice.Message.Role,
+				Content:   choice.Message.Content,
+				ToolCalls: nil,
+			},
+			FinishReason: choice.FinishReason,
+		})
+	}
+	return &chatCompletionResponse{
+		Id:                hunyuanResponse.Response.Id,
+		Created:           time.Now().UnixMilli() / 1000,
+		Model:             ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
+		SystemFingerprint: "",
+		Object:            objectChatCompletion,
+		Choices:           choices,
+		Usage: usage{
+			PromptTokens:     hunyuanResponse.Response.Usage.PromptTokens,
+			CompletionTokens: hunyuanResponse.Response.Usage.CompletionTokens,
+			TotalTokens:      hunyuanResponse.Response.Usage.TotalTokens,
+		},
+	}
+}
+
+func Sha256hex(s string) string {
+	b := sha256.Sum256([]byte(s))
+	return hex.EncodeToString(b[:])
+}
+
+func Hmacsha256(s, key string) string {
+	hashed := hmac.New(sha256.New, []byte(key))
+	hashed.Write([]byte(s))
+	return string(hashed.Sum(nil))
+}
+
+/**
+ * @param secretId 秘钥id
+ * @param secretKey 秘钥
+ * @param timestamp 时间戳
+ * @param host 目标域名
+ * @param action 请求动作
+ * @param payload 请求体
+ * @return 签名
+ */
+func GetTC3Authorizationcode(secretId string, secretKey string, timestamp int64, host string, action string, payload string) string {
+	algorithm := "TC3-HMAC-SHA256"
+	service := "hunyuan" // 注意，必须和域名中的产品名保持一致
+
+	// step 1: build canonical request string
+	httpRequestMethod := "POST"
+	canonicalURI := "/"
+	canonicalQueryString := ""
+	canonicalHeaders := fmt.Sprintf("content-type:%s\nhost:%s\nx-tc-action:%s\n",
+		"application/json", host, strings.ToLower(action))
+	signedHeaders := "content-type;host;x-tc-action"
+
+	// fmt.Println("payload is: %s", payload)
+	hashedRequestPayload := Sha256hex(payload)
+	canonicalRequest := fmt.Sprintf("%s\n%s\n%s\n%s\n%s\n%s",
+		httpRequestMethod,
+		canonicalURI,
+		canonicalQueryString,
+		canonicalHeaders,
+		signedHeaders,
+		hashedRequestPayload)
+	// fmt.Println(canonicalRequest)
+
+	// step 2: build string to sign
+	date := time.Unix(timestamp, 0).UTC().Format("2006-01-02")
+	credentialScope := fmt.Sprintf("%s/%s/tc3_request", date, service)
+	hashedCanonicalRequest := Sha256hex(canonicalRequest)
+	string2sign := fmt.Sprintf("%s\n%d\n%s\n%s",
+		algorithm,
+		timestamp,
+		credentialScope,
+		hashedCanonicalRequest)
+	// fmt.Println(string2sign)
+
+	// step 3: sign string
+	secretDate := Hmacsha256(date, "TC3"+secretKey)
+	secretService := Hmacsha256(service, secretDate)
+	secretSigning := Hmacsha256("tc3_request", secretService)
+	signature := hex.EncodeToString([]byte(Hmacsha256(string2sign, secretSigning)))
+	// fmt.Println(signature)
+
+	// step 4: build authorization
+	authorization := fmt.Sprintf("%s Credential=%s/%s, SignedHeaders=%s, Signature=%s",
+		algorithm,
+		secretId,
+		credentialScope,
+		signedHeaders,
+		signature)
+
+	// curl := fmt.Sprintf(`curl -X POST https://%s \
+	// 	-H "Authorization: %s" \
+	// 	-H "Content-Type: application/json" \
+	// 	-H "Host: %s" -H "X-TC-Action: %s" \
+	// 	-H "X-TC-Timestamp: %d" \
+	// 	-H "X-TC-Version: 2023-09-01" \
+	// 	-d '%s'`, host, authorization, host, action, timestamp, payload)
+	// fmt.Println(curl)
+	return authorization
+}
--- a/plugins/wasm-go/extensions/ai-proxy/provider/minimax.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/minimax.go
@@ -0,0 +1,476 @@
+package provider
+
+import (
+	"encoding/json"
+	"errors"
+	"fmt"
+	"strings"
+
+	"github.com/alibaba/higress/plugins/wasm-go/extensions/ai-proxy/util"
+	"github.com/alibaba/higress/plugins/wasm-go/pkg/wrapper"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm/types"
+)
+
+// minimaxProvider is the provider for minimax service.
+
+const (
+	minimaxDomain = "api.minimax.chat"
+	// minimaxChatCompletionV2Path 接口请求响应格式与OpenAI相同
+	// 接口文档: https://platform.minimaxi.com/document/guides/chat-model/V2?id=65e0736ab2845de20908e2dd
+	minimaxChatCompletionV2Path = "/v1/text/chatcompletion_v2"
+	// minimaxChatCompletionProPath 接口请求响应格式与OpenAI不同
+	// 接口文档: https://platform.minimaxi.com/document/guides/chat-model/pro/api?id=6569c85948bc7b684b30377e
+	minimaxChatCompletionProPath = "/v1/text/chatcompletion_pro"
+
+	senderTypeUser string = "USER" // 用户发送的内容
+	senderTypeBot  string = "BOT"  // 模型生成的内容
+
+	// 默认机器人设置
+	defaultBotName           string = "MM智能助理"
+	defaultBotSettingContent string = "MM智能助理是一款由MiniMax自研的，没有调用其他产品的接口的大型语言模型。MiniMax是一家中国科技公司，一直致力于进行大模型相关的研究。"
+	defaultSenderName        string = "小明"
+)
+
+// chatCompletionProModels 这些模型对应接口为ChatCompletion Pro
+var chatCompletionProModels = map[string]struct{}{
+	"abab6.5-chat":  {},
+	"abab6.5s-chat": {},
+	"abab5.5s-chat": {},
+	"abab5.5-chat":  {},
+}
+
+type minimaxProviderInitializer struct {
+}
+
+func (m *minimaxProviderInitializer) ValidateConfig(config ProviderConfig) error {
+	// 如果存在模型对应接口为ChatCompletion Pro必须配置minimaxGroupId
+	if len(config.modelMapping) > 0 && config.minimaxGroupId == "" {
+		for _, minimaxModel := range config.modelMapping {
+			if _, exists := chatCompletionProModels[minimaxModel]; exists {
+				return errors.New(fmt.Sprintf("missing minimaxGroupId in provider config when %s model is provided", minimaxModel))
+			}
+		}
+	}
+	if config.apiTokens == nil || len(config.apiTokens) == 0 {
+		return errors.New("no apiToken found in provider config")
+	}
+	return nil
+}
+
+func (m *minimaxProviderInitializer) CreateProvider(config ProviderConfig) (Provider, error) {
+	return &minimaxProvider{
+		config:       config,
+		contextCache: createContextCache(&config),
+	}, nil
+}
+
+type minimaxProvider struct {
+	config       ProviderConfig
+	contextCache *contextCache
+}
+
+func (m *minimaxProvider) GetProviderType() string {
+	return providerTypeMinimax
+}
+
+func (m *minimaxProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+	_ = util.OverwriteRequestHost(minimaxDomain)
+	_ = util.OverwriteRequestAuthorization("Bearer " + m.config.GetRandomToken())
+	_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
+
+	// Delay the header processing to allow changing streaming mode in OnRequestBody
+	return types.HeaderStopIteration, nil
+}
+
+func (m *minimaxProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+	// 解析并映射模型,设置上下文
+	model, err := m.parseModel(body)
+	if err != nil {
+		return types.ActionContinue, err
+	}
+	ctx.SetContext(ctxKeyOriginalRequestModel, model)
+	mappedModel := getMappedModel(model, m.config.modelMapping, log)
+	if mappedModel == "" {
+		return types.ActionContinue, errors.New("model becomes empty after applying the configured mapping")
+	}
+	ctx.SetContext(ctxKeyFinalRequestModel, mappedModel)
+	_, ok := chatCompletionProModels[mappedModel]
+	if ok {
+		// 使用ChatCompletion Pro接口
+		return m.handleRequestBodyByChatCompletionPro(body, log)
+	} else {
+		// 使用ChatCompletion v2接口
+		return m.handleRequestBodyByChatCompletionV2(body, log)
+	}
+}
+
+// handleRequestBodyByChatCompletionPro 使用ChatCompletion Pro接口处理请求体
+func (m *minimaxProvider) handleRequestBodyByChatCompletionPro(body []byte, log wrapper.Log) (types.Action, error) {
+	// 使用minimax接口协议
+	if m.config.protocol == protocolOriginal {
+		request := &minimaxChatCompletionV2Request{}
+		if err := json.Unmarshal(body, request); err != nil {
+			return types.ActionContinue, fmt.Errorf("unable to unmarshal request: %v", err)
+		}
+		if request.Model == "" {
+			return types.ActionContinue, errors.New("request model is empty")
+		}
+		// 根据模型重写requestPath
+		if m.config.minimaxGroupId == "" {
+			return types.ActionContinue, errors.New(fmt.Sprintf("missing minimaxGroupId in provider config when use %s model ", request.Model))
+		}
+		_ = util.OverwriteRequestPath(fmt.Sprintf("%s?GroupId=%s", minimaxChatCompletionProPath, m.config.minimaxGroupId))
+
+		if m.config.context == nil {
+			return types.ActionContinue, nil
+		}
+
+		err := m.contextCache.GetContent(func(content string, err error) {
+			defer func() {
+				_ = proxywasm.ResumeHttpRequest()
+			}()
+
+			if err != nil {
+				log.Errorf("failed to load context file: %v", err)
+				_ = util.SendResponse(500, "ai-proxy.minimax.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+			}
+			m.setBotSettings(request, content)
+			if err := replaceJsonRequestBody(request, log); err != nil {
+				_ = util.SendResponse(500, "ai-proxy.minimax.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+			}
+		}, log)
+		if err == nil {
+			return types.ActionPause, nil
+		}
+		return types.ActionContinue, err
+	}
+
+	request := &chatCompletionRequest{}
+	if err := decodeChatCompletionRequest(body, request); err != nil {
+		return types.ActionContinue, err
+	}
+
+	// 映射模型重写requestPath
+	request.Model = getMappedModel(request.Model, m.config.modelMapping, log)
+	_ = util.OverwriteRequestPath(fmt.Sprintf("%s?GroupId=%s", minimaxChatCompletionProPath, m.config.minimaxGroupId))
+
+	if m.config.context == nil {
+		minimaxRequest := m.buildMinimaxChatCompletionV2Request(request, "")
+		return types.ActionContinue, replaceJsonRequestBody(minimaxRequest, log)
+	}
+
+	err := m.contextCache.GetContent(func(content string, err error) {
+		defer func() {
+			_ = proxywasm.ResumeHttpRequest()
+		}()
+		if err != nil {
+			log.Errorf("failed to load context file: %v", err)
+			_ = util.SendResponse(500, "ai-proxy.minimax.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+		}
+		minimaxRequest := m.buildMinimaxChatCompletionV2Request(request, content)
+		if err := replaceJsonRequestBody(minimaxRequest, log); err != nil {
+			_ = util.SendResponse(500, "ai-proxy.minimax.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace Request body: %v", err))
+		}
+	}, log)
+	if err == nil {
+		return types.ActionPause, nil
+	}
+	return types.ActionContinue, err
+}
+
+// handleRequestBodyByChatCompletionV2 使用ChatCompletion v2接口处理请求体
+func (m *minimaxProvider) handleRequestBodyByChatCompletionV2(body []byte, log wrapper.Log) (types.Action, error) {
+	request := &chatCompletionRequest{}
+	if err := decodeChatCompletionRequest(body, request); err != nil {
+		return types.ActionContinue, err
+	}
+
+	// 映射模型重写requestPath
+	request.Model = getMappedModel(request.Model, m.config.modelMapping, log)
+	_ = util.OverwriteRequestPath(minimaxChatCompletionV2Path)
+
+	if m.contextCache == nil {
+		return types.ActionContinue, replaceJsonRequestBody(request, log)
+	}
+
+	err := m.contextCache.GetContent(func(content string, err error) {
+		defer func() {
+			_ = proxywasm.ResumeHttpRequest()
+		}()
+		if err != nil {
+			log.Errorf("failed to load context file: %v", err)
+			_ = util.SendResponse(500, "ai-proxy.minimax.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+		}
+		insertContextMessage(request, content)
+		if err := replaceJsonRequestBody(request, log); err != nil {
+			_ = util.SendResponse(500, "ai-proxy.minimax.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+		}
+	}, log)
+	if err == nil {
+		return types.ActionPause, nil
+	}
+	return types.ActionContinue, err
+}
+
+func (m *minimaxProvider) OnResponseHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	// 使用minimax接口协议,跳过OnStreamingResponseBody()和OnResponseBody()
+	if m.config.protocol == protocolOriginal {
+		ctx.DontReadResponseBody()
+		return types.ActionContinue, nil
+	}
+	// 模型对应接口为ChatCompletion v2,跳过OnStreamingResponseBody()和OnResponseBody()
+	model := ctx.GetStringContext(ctxKeyFinalRequestModel, "")
+	if model != "" {
+		_, ok := chatCompletionProModels[model]
+		if !ok {
+			ctx.DontReadResponseBody()
+			return types.ActionContinue, nil
+		}
+	}
+	_ = proxywasm.RemoveHttpResponseHeader("Content-Length")
+	return types.ActionContinue, nil
+}
+
+// OnStreamingResponseBody 只处理使用OpenAI协议 且 模型对应接口为ChatCompletion Pro的流式响应
+func (m *minimaxProvider) OnStreamingResponseBody(ctx wrapper.HttpContext, name ApiName, chunk []byte, isLastChunk bool, log wrapper.Log) ([]byte, error) {
+	if isLastChunk || len(chunk) == 0 {
+		return nil, nil
+	}
+	// sample event response:
+	// data: {"created":1689747645,"model":"abab6.5s-chat","reply":"","choices":[{"messages":[{"sender_type":"BOT","sender_name":"MM智能助理","text":"am from China."}]}],"output_sensitive":false}
+
+	// sample end event response:
+	// data: {"created":1689747645,"model":"abab6.5s-chat","reply":"I am from China.","choices":[{"finish_reason":"stop","messages":[{"sender_type":"BOT","sender_name":"MM智能助理","text":"I am from China."}]}],"usage":{"total_tokens":187},"input_sensitive":false,"output_sensitive":false,"id":"0106b3bc9fd844a9f3de1aa06004e2ab","base_resp":{"status_code":0,"status_msg":""}}
+	responseBuilder := &strings.Builder{}
+	lines := strings.Split(string(chunk), "\n")
+	for _, data := range lines {
+		if len(data) < 6 {
+			// ignore blank line or wrong format
+			continue
+		}
+		data = data[6:]
+		var minimaxResp minimaxChatCompletionV2Resp
+		if err := json.Unmarshal([]byte(data), &minimaxResp); err != nil {
+			log.Errorf("unable to unmarshal minimax response: %v", err)
+			continue
+		}
+		response := m.responseV2ToOpenAI(&minimaxResp)
+		responseBody, err := json.Marshal(response)
+		if err != nil {
+			log.Errorf("unable to marshal response: %v", err)
+			return nil, err
+		}
+		m.appendResponse(responseBuilder, string(responseBody))
+	}
+	modifiedResponseChunk := responseBuilder.String()
+	log.Debugf("=== modified response chunk: %s", modifiedResponseChunk)
+	return []byte(modifiedResponseChunk), nil
+}
+
+// OnResponseBody 只处理使用OpenAI协议 且 模型对应接口为ChatCompletion Pro的流式响应
+func (m *minimaxProvider) OnResponseBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	minimaxResp := &minimaxChatCompletionV2Resp{}
+	if err := json.Unmarshal(body, minimaxResp); err != nil {
+		return types.ActionContinue, fmt.Errorf("unable to unmarshal minimax response: %v", err)
+	}
+	if minimaxResp.BaseResp.StatusCode != 0 {
+		return types.ActionContinue, fmt.Errorf("minimax response error, error_code: %d, error_message: %s", minimaxResp.BaseResp.StatusCode, minimaxResp.BaseResp.StatusMsg)
+	}
+	response := m.responseV2ToOpenAI(minimaxResp)
+	return types.ActionContinue, replaceJsonResponseBody(response, log)
+}
+
+// minimaxChatCompletionV2Request 表示ChatCompletion V2请求的结构体
+type minimaxChatCompletionV2Request struct {
+	Model             string                  `json:"model"`
+	Stream            bool                    `json:"stream,omitempty"`
+	TokensToGenerate  int64                   `json:"tokens_to_generate,omitempty"`
+	Temperature       float64                 `json:"temperature,omitempty"`
+	TopP              float64                 `json:"top_p,omitempty"`
+	MaskSensitiveInfo bool                    `json:"mask_sensitive_info"` // 是否开启隐私信息打码,默认true
+	Messages          []minimaxMessage        `json:"messages"`
+	BotSettings       []minimaxBotSetting     `json:"bot_setting"`
+	ReplyConstraints  minimaxReplyConstraints `json:"reply_constraints"`
+}
+
+// minimaxMessage 表示对话中的消息
+type minimaxMessage struct {
+	SenderType string `json:"sender_type"`
+	SenderName string `json:"sender_name"`
+	Text       string `json:"text"`
+}
+
+// minimaxBotSetting 表示机器人的设置
+type minimaxBotSetting struct {
+	BotName string `json:"bot_name"`
+	Content string `json:"content"`
+}
+
+// minimaxReplyConstraints 表示模型回复要求
+type minimaxReplyConstraints struct {
+	SenderType string `json:"sender_type"`
+	SenderName string `json:"sender_name"`
+}
+
+// minimaxChatCompletionV2Resp Minimax Chat Completion V2响应结构体
+type minimaxChatCompletionV2Resp struct {
+	Created             int64           `json:"created"`
+	Model               string          `json:"model"`
+	Reply               string          `json:"reply"`
+	InputSensitive      bool            `json:"input_sensitive,omitempty"`
+	InputSensitiveType  int64           `json:"input_sensitive_type,omitempty"`
+	OutputSensitive     bool            `json:"output_sensitive,omitempty"`
+	OutputSensitiveType int64           `json:"output_sensitive_type,omitempty"`
+	Choices             []minimaxChoice `json:"choices,omitempty"`
+	Usage               minimaxUsage    `json:"usage,omitempty"`
+	Id                  string          `json:"id"`
+	BaseResp            minimaxBaseResp `json:"base_resp"`
+}
+
+// minimaxBaseResp 包含错误状态码和详情
+type minimaxBaseResp struct {
+	StatusCode int64  `json:"status_code"`
+	StatusMsg  string `json:"status_msg"`
+}
+
+// minimaxChoice 结果选项
+type minimaxChoice struct {
+	Messages     []minimaxMessage `json:"messages"`
+	Index        int64            `json:"index"`
+	FinishReason string           `json:"finish_reason"`
+}
+
+// minimaxUsage 令牌使用情况
+type minimaxUsage struct {
+	TotalTokens int64 `json:"total_tokens"`
+}
+
+func (m *minimaxProvider) parseModel(body []byte) (string, error) {
+	var tempMap map[string]interface{}
+	if err := json.Unmarshal(body, &tempMap); err != nil {
+		return "", err
+	}
+	model, ok := tempMap["model"].(string)
+	if !ok {
+		return "", errors.New("missing model in chat completion request")
+	}
+	return model, nil
+}
+
+func (m *minimaxProvider) setBotSettings(request *minimaxChatCompletionV2Request, botSettingContent string) {
+	if len(request.BotSettings) == 0 {
+		request.BotSettings = []minimaxBotSetting{
+			{
+				BotName: defaultBotName,
+				Content: func() string {
+					if botSettingContent != "" {
+						return botSettingContent
+					}
+					return defaultBotSettingContent
+				}(),
+			},
+		}
+	} else if botSettingContent != "" {
+		newSetting := minimaxBotSetting{
+			BotName: request.BotSettings[0].BotName,
+			Content: botSettingContent,
+		}
+		request.BotSettings = append([]minimaxBotSetting{newSetting}, request.BotSettings...)
+	}
+}
+
+func (m *minimaxProvider) buildMinimaxChatCompletionV2Request(request *chatCompletionRequest, botSettingContent string) *minimaxChatCompletionV2Request {
+	var messages []minimaxMessage
+	var botSetting []minimaxBotSetting
+	var botName string
+
+	determineName := func(name string, defaultName string) string {
+		if name != "" {
+			return name
+		}
+		return defaultName
+	}
+
+	for _, message := range request.Messages {
+		switch message.Role {
+		case roleSystem:
+			botName = determineName(message.Name, defaultBotName)
+			botSetting = append(botSetting, minimaxBotSetting{
+				BotName: botName,
+				Content: message.Content,
+			})
+		case roleAssistant:
+			messages = append(messages, minimaxMessage{
+				SenderType: senderTypeBot,
+				SenderName: determineName(message.Name, defaultBotName),
+				Text:       message.Content,
+			})
+		case roleUser:
+			messages = append(messages, minimaxMessage{
+				SenderType: senderTypeUser,
+				SenderName: determineName(message.Name, defaultSenderName),
+				Text:       message.Content,
+			})
+		}
+	}
+
+	replyConstraints := minimaxReplyConstraints{
+		SenderType: senderTypeBot,
+		SenderName: determineName(botName, defaultBotName),
+	}
+	result := &minimaxChatCompletionV2Request{
+		Model:             request.Model,
+		Stream:            request.Stream,
+		TokensToGenerate:  int64(request.MaxTokens),
+		Temperature:       request.Temperature,
+		TopP:              request.TopP,
+		MaskSensitiveInfo: true,
+		Messages:          messages,
+		BotSettings:       botSetting,
+		ReplyConstraints:  replyConstraints,
+	}
+
+	m.setBotSettings(result, botSettingContent)
+	return result
+}
+
+func (m *minimaxProvider) responseV2ToOpenAI(response *minimaxChatCompletionV2Resp) *chatCompletionResponse {
+	var choices []chatCompletionChoice
+	messageIndex := 0
+	for _, choice := range response.Choices {
+		for _, message := range choice.Messages {
+			message := &chatMessage{
+				Name:    message.SenderName,
+				Role:    roleAssistant,
+				Content: message.Text,
+			}
+			choices = append(choices, chatCompletionChoice{
+				FinishReason: choice.FinishReason,
+				Index:        messageIndex,
+				Message:      message,
+			})
+			messageIndex++
+		}
+	}
+	return &chatCompletionResponse{
+		Id:      response.Id,
+		Object:  objectChatCompletion,
+		Created: response.Created,
+		Model:   response.Model,
+		Choices: choices,
+		Usage: usage{
+			TotalTokens: int(response.Usage.TotalTokens),
+		},
+	}
+}
+
+func (m *minimaxProvider) appendResponse(responseBuilder *strings.Builder, responseBody string) {
+	responseBuilder.WriteString(fmt.Sprintf("%s %s\n\n", streamDataItemKey, responseBody))
+}
--- a/plugins/wasm-go/extensions/ai-proxy/provider/model.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/model.go
@@ -30,6 +30,7 @@ type chatCompletionRequest struct {
 	Tools            []tool         `json:"tools,omitempty"`
 	ToolChoice       *toolChoice    `json:"tool_choice,omitempty"`
 	User             string         `json:"user,omitempty"`
+	Stop             []string       `json:"stop,omitempty"`
 }

 type streamOptions struct {
@@ -59,7 +60,7 @@ type chatCompletionResponse struct {
 	Model             string                 `json:"model,omitempty"`
 	SystemFingerprint string                 `json:"system_fingerprint,omitempty"`
 	Object            string                 `json:"object,omitempty"`
-	Usage             chatCompletionUsage    `json:"usage,omitempty"`
+	Usage             usage                  `json:"usage,omitempty"`
 }

 type chatCompletionChoice struct {
@@ -69,7 +70,7 @@ type chatCompletionChoice struct {
 	FinishReason string       `json:"finish_reason,omitempty"`
 }

-type chatCompletionUsage struct {
+type usage struct {
 	PromptTokens     int `json:"prompt_tokens,omitempty"`
 	CompletionTokens int `json:"completion_tokens,omitempty"`
 	TotalTokens      int `json:"total_tokens,omitempty"`
@@ -139,3 +140,43 @@ func (e *streamEvent) setValue(key, value string) {
 		}
 	}
 }
+
+type embeddingsRequest struct {
+	Input          interface{} `json:"input"`
+	Model          string      `json:"model"`
+	EncodingFormat string      `json:"encoding_format,omitempty"`
+	Dimensions     int         `json:"dimensions,omitempty"`
+	User           string      `json:"user,omitempty"`
+}
+
+type embeddingsResponse struct {
+	Object string      `json:"object"`
+	Data   []embedding `json:"data"`
+	Model  string      `json:"model"`
+	Usage  usage       `json:"usage"`
+}
+
+type embedding struct {
+	Object    string    `json:"object"`
+	Index     int       `json:"index"`
+	Embedding []float64 `json:"embedding"`
+}
+
+func (r embeddingsRequest) ParseInput() []string {
+	if r.Input == nil {
+		return nil
+	}
+	var input []string
+	switch r.Input.(type) {
+	case string:
+		input = []string{r.Input.(string)}
+	case []any:
+		input = make([]string, 0, len(r.Input.([]any)))
+		for _, item := range r.Input.([]any) {
+			if str, ok := item.(string); ok {
+				input = append(input, str)
+			}
+		}
+	}
+	return input
+}
--- a/plugins/wasm-go/extensions/ai-proxy/provider/moonshot.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/moonshot.go
@@ -26,6 +26,9 @@ func (m *moonshotProviderInitializer) ValidateConfig(config ProviderConfig) erro
 	if config.moonshotFileId != "" && config.context != nil {
 		return errors.New("moonshotFileId and context cannot be configured at the same time")
 	}
+	if config.apiTokens == nil || len(config.apiTokens) == 0 {
+		return errors.New("no apiToken found in provider config")
+	}
 	return nil
 }

@@ -57,7 +60,7 @@ func (m *moonshotProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName Api
 	}
 	_ = util.OverwriteRequestPath(moonshotChatCompletionPath)
 	_ = util.OverwriteRequestHost(moonshotDomain)
-	_ = proxywasm.ReplaceHttpRequestHeader("Authorization", "Bearer "+m.config.GetRandomToken())
+	_ = util.OverwriteRequestAuthorization("Bearer " + m.config.GetRandomToken())
 	_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
 	return types.ActionContinue, nil
 }
@@ -92,12 +95,12 @@ func (m *moonshotProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiNam
 		}()
 		if err != nil {
 			log.Errorf("failed to load context file: %v", err)
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.moonshot.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
 			return
 		}
 		err = m.performChatCompletion(ctx, content, request, log)
 		if err != nil {
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to perform chat completion: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.moonshot.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to perform chat completion: %v", err))
 		}
 	}, log)
 	if err == nil {
--- a/plugins/wasm-go/extensions/ai-proxy/provider/ollama.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/ollama.go
@@ -68,7 +68,7 @@ func (m *ollamaProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName,
 	if m.config.modelMapping == nil && m.contextCache == nil {
 		return types.ActionContinue, nil
 	}
-	
+
 	request := &chatCompletionRequest{}
 	if err := decodeChatCompletionRequest(body, request); err != nil {
 		return types.ActionContinue, err
@@ -83,7 +83,7 @@ func (m *ollamaProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName,
 		return types.ActionContinue, errors.New("model becomes empty after applying the configured mapping")
 	}
 	request.Model = mappedModel
-	
+
 	if m.contextCache != nil {
 		err := m.contextCache.GetContent(func(content string, err error) {
 			defer func() {
@@ -91,11 +91,11 @@ func (m *ollamaProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName,
 			}()
 			if err != nil {
 				log.Errorf("failed to load context file: %v", err)
-				_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+				_ = util.SendResponse(500, "ai-proxy.ollama.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
 			}
 			insertContextMessage(request, content)
 			if err := replaceJsonRequestBody(request, log); err != nil {
-				_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+				_ = util.SendResponse(500, "ai-proxy.ollama.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
 			}
 		}, log)
 		if err == nil {
@@ -105,7 +105,7 @@ func (m *ollamaProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName,
 		}
 	} else {
 		if err := replaceJsonRequestBody(request, log); err != nil {
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.ollama.transform_body_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
 			return types.ActionContinue, err
 		}
 		_ = proxywasm.ResumeHttpRequest()
--- a/plugins/wasm-go/extensions/ai-proxy/provider/openai.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/openai.go
@@ -2,6 +2,7 @@ package provider

 import (
 	"fmt"
+	"strings"

 	"github.com/alibaba/higress/plugins/wasm-go/extensions/ai-proxy/util"
 	"github.com/alibaba/higress/plugins/wasm-go/pkg/wrapper"
@@ -12,8 +13,9 @@ import (
 // openaiProvider is the provider for OpenAI service.

 const (
-	openaiDomain             = "api.openai.com"
-	openaiChatCompletionPath = "/v1/chat/completions"
+	defaultOpenaiDomain             = "api.openai.com"
+	defaultOpenaiChatCompletionPath = "/v1/chat/completions"
+	defaultOpenaiEmbeddingsPath     = "/v1/chat/embeddings"
 )

 type openaiProviderInitializer struct {
@@ -24,14 +26,29 @@ func (m *openaiProviderInitializer) ValidateConfig(config ProviderConfig) error
 }

 func (m *openaiProviderInitializer) CreateProvider(config ProviderConfig) (Provider, error) {
+	if config.openaiCustomUrl == "" {
+		return &openaiProvider{
+			config:       config,
+			contextCache: createContextCache(&config),
+		}, nil
+	}
+	customUrl := strings.TrimPrefix(strings.TrimPrefix(config.openaiCustomUrl, "http://"), "https://")
+	pairs := strings.SplitN(customUrl, "/", 2)
+	if len(pairs) != 2 {
+		return nil, fmt.Errorf("invalid openaiCustomUrl:%s", config.openaiCustomUrl)
+	}
 	return &openaiProvider{
 		config:       config,
+		customDomain: pairs[0],
+		customPath:   "/" + pairs[1],
 		contextCache: createContextCache(&config),
 	}, nil
 }

 type openaiProvider struct {
 	config       ProviderConfig
+	customDomain string
+	customPath   string
 	contextCache *contextCache
 }

@@ -40,44 +57,63 @@ func (m *openaiProvider) GetProviderType() string {
 }

 func (m *openaiProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
-	if apiName != ApiNameChatCompletion {
-		return types.ActionContinue, errUnsupportedApiName
-	}
-	_ = util.OverwriteRequestPath(openaiChatCompletionPath)
-	_ = util.OverwriteRequestHost(openaiDomain)
-	_ = proxywasm.ReplaceHttpRequestHeader("Authorization", "Bearer "+m.config.GetRandomToken())
-
-	if m.contextCache == nil {
-		ctx.DontReadRequestBody()
+	if m.customPath == "" {
+		switch apiName {
+		case ApiNameChatCompletion:
+			_ = util.OverwriteRequestPath(defaultOpenaiChatCompletionPath)
+		case ApiNameEmbeddings:
+			ctx.DontReadRequestBody()
+			_ = util.OverwriteRequestPath(defaultOpenaiEmbeddingsPath)
+		}
 	} else {
-		_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
+		_ = util.OverwriteRequestPath(m.customPath)
 	}
-
+	if m.customDomain == "" {
+		_ = util.OverwriteRequestHost(defaultOpenaiDomain)
+	} else {
+		_ = util.OverwriteRequestHost(m.customDomain)
+	}
+	if len(m.config.apiTokens) > 0 {
+		_ = util.OverwriteRequestAuthorization("Bearer " + m.config.GetRandomToken())
+	}
+	_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
 	return types.ActionContinue, nil
 }

 func (m *openaiProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
 	if apiName != ApiNameChatCompletion {
-		return types.ActionContinue, errUnsupportedApiName
-	}
-	if m.contextCache == nil {
+		// We don't need to process the request body for other APIs.
 		return types.ActionContinue, nil
 	}
 	request := &chatCompletionRequest{}
 	if err := decodeChatCompletionRequest(body, request); err != nil {
 		return types.ActionContinue, err
 	}
+	if request.Stream {
+		// For stream requests, we need to include usage in the response.
+		if request.StreamOptions == nil {
+			request.StreamOptions = &streamOptions{IncludeUsage: true}
+		} else if !request.StreamOptions.IncludeUsage {
+			request.StreamOptions.IncludeUsage = true
+		}
+	}
+	if m.contextCache == nil {
+		if err := replaceJsonRequestBody(request, log); err != nil {
+			_ = util.SendResponse(500, "ai-proxy.openai.set_include_usage_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+		}
+		return types.ActionContinue, nil
+	}
 	err := m.contextCache.GetContent(func(content string, err error) {
 		defer func() {
 			_ = proxywasm.ResumeHttpRequest()
 		}()
 		if err != nil {
 			log.Errorf("failed to load context file: %v", err)
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.openai.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
 		}
 		insertContextMessage(request, content)
 		if err := replaceJsonRequestBody(request, log); err != nil {
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.openai.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
 		}
 	}, log)
 	if err == nil {
--- a/plugins/wasm-go/extensions/ai-proxy/provider/provider.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/provider.go
@@ -3,6 +3,7 @@ package provider
 import (
 	"errors"
 	"math/rand"
+	"strings"

 	"github.com/alibaba/higress/plugins/wasm-go/pkg/wrapper"
 	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm/types"
@@ -14,24 +15,39 @@ type Pointcut string

 const (
 	ApiNameChatCompletion ApiName = "chatCompletion"
+	ApiNameEmbeddings     ApiName = "embeddings"

-	providerTypeMoonshot = "moonshot"
-	providerTypeAzure    = "azure"
-	providerTypeQwen     = "qwen"
-	providerTypeOpenAI   = "openai"
-	providerTypeGroq     = "groq"
-	providerTypeBaichuan = "baichuan"
-	providerTypeYi       = "yi"
-	providerTypeDeepSeek = "deepseek"
-	providerTypeZhipuAi  = "zhipuai"
-	providerTypeOllama   = "ollama"
+	providerTypeMoonshot   = "moonshot"
+	providerTypeAzure      = "azure"
+	providerTypeQwen       = "qwen"
+	providerTypeOpenAI     = "openai"
+	providerTypeGroq       = "groq"
+	providerTypeBaichuan   = "baichuan"
+	providerTypeYi         = "yi"
+	providerTypeDeepSeek   = "deepseek"
+	providerTypeZhipuAi    = "zhipuai"
+	providerTypeOllama     = "ollama"
+	providerTypeClaude     = "claude"
+	providerTypeBaidu      = "baidu"
+	providerTypeHunyuan    = "hunyuan"
+	providerTypeStepfun    = "stepfun"
+	providerTypeMinimax    = "minimax"
+	providerTypeCloudflare = "cloudflare"
+	providerTypeSpark      = "spark"
+	providerTypeGemini     = "gemini"

 	protocolOpenAI   = "openai"
 	protocolOriginal = "original"

-	roleSystem = "system"
+	roleSystem    = "system"
+	roleAssistant = "assistant"
+	roleUser      = "user"
+
+	finishReasonStop   = "stop"
+	finishReasonLength = "length"

 	ctxKeyIncrementalStreaming = "incrementalStreaming"
+	ctxKeyApiName              = "apiKey"
 	ctxKeyStreamingBody        = "streamingBody"
 	ctxKeyOriginalRequestModel = "originalRequestModel"
 	ctxKeyFinalRequestModel    = "finalRequestModel"
@@ -54,16 +70,24 @@ var (
 	errUnsupportedApiName = errors.New("unsupported API name")

 	providerInitializers = map[string]providerInitializer{
-		providerTypeMoonshot: &moonshotProviderInitializer{},
-		providerTypeAzure:    &azureProviderInitializer{},
-		providerTypeQwen:     &qwenProviderInitializer{},
-		providerTypeOpenAI:   &openaiProviderInitializer{},
-		providerTypeGroq:     &groqProviderInitializer{},
-		providerTypeBaichuan: &baichuanProviderInitializer{},
-		providerTypeYi:       &yiProviderInitializer{},
-		providerTypeDeepSeek: &deepseekProviderInitializer{},
-		providerTypeZhipuAi:  &zhipuAiProviderInitializer{},
-		providerTypeOllama:   &ollamaProviderInitializer{},
+		providerTypeMoonshot:   &moonshotProviderInitializer{},
+		providerTypeAzure:      &azureProviderInitializer{},
+		providerTypeQwen:       &qwenProviderInitializer{},
+		providerTypeOpenAI:     &openaiProviderInitializer{},
+		providerTypeGroq:       &groqProviderInitializer{},
+		providerTypeBaichuan:   &baichuanProviderInitializer{},
+		providerTypeYi:         &yiProviderInitializer{},
+		providerTypeDeepSeek:   &deepseekProviderInitializer{},
+		providerTypeZhipuAi:    &zhipuAiProviderInitializer{},
+		providerTypeOllama:     &ollamaProviderInitializer{},
+		providerTypeClaude:     &claudeProviderInitializer{},
+		providerTypeBaidu:      &baiduProviderInitializer{},
+		providerTypeHunyuan:    &hunyuanProviderInitializer{},
+		providerTypeStepfun:    &stepfunProviderInitializer{},
+		providerTypeMinimax:    &minimaxProviderInitializer{},
+		providerTypeCloudflare: &cloudflareProviderInitializer{},
+		providerTypeSpark:      &sparkProviderInitializer{},
+		providerTypeGemini:     &geminiProviderInitializer{},
 	}
 )

@@ -93,7 +117,7 @@ type ResponseBodyHandler interface {

 type ProviderConfig struct {
 	// @Title zh-CN AI服务提供商
-	// @Description zh-CN AI服务提供商类型，目前支持的取值为："moonshot"、"qwen"、"openai"、"azure"、"baichuan"、"yi"、"zhipuai"、"ollama"
+	// @Description zh-CN AI服务提供商类型
 	typ string `required:"true" yaml:"type" json:"type"`
 	// @Title zh-CN API Tokens
 	// @Description zh-CN 在请求AI服务时用于认证的API Token列表。不同的AI服务提供商可能有不同的名称。部分供应商只支持配置一个API Token（如Azure OpenAI）。
@@ -101,6 +125,9 @@ type ProviderConfig struct {
 	// @Title zh-CN 请求超时
 	// @Description zh-CN 请求AI服务的超时时间，单位为毫秒。默认值为120000，即2分钟
 	timeout uint32 `required:"false" yaml:"timeout" json:"timeout"`
+	// @Title zh-CN 基于OpenAI协议的自定义后端URL
+	// @Description zh-CN 仅适用于支持 openai 协议的服务。
+	openaiCustomUrl string `required:"false" yaml:"openaiCustomUrl" json:"openaiCustomUrl"`
 	// @Title zh-CN Moonshot File ID
 	// @Description zh-CN 仅适用于Moonshot AI服务。Moonshot AI服务的文件ID，其内容用于补充AI请求上下文
 	moonshotFileId string `required:"false" yaml:"moonshotFileId" json:"moonshotFileId"`
@@ -119,6 +146,15 @@ type ProviderConfig struct {
 	// @Title zh-CN Ollama Server Port
 	// @Description zh-CN 仅适用于 Ollama 服务。Ollama 服务器的端口号。
 	ollamaServerPort uint32 `required:"false" yaml:"ollamaServerPort" json:"ollamaServerPort"`
+	// @Title zh-CN hunyuan api key for authorization
+	// @Description zh-CN 仅适用于Hun Yuan AI服务鉴权，API key/id 参考：https://cloud.tencent.com/document/api/1729/101843#Golang
+	hunyuanAuthKey string `required:"false" yaml:"hunyuanAuthKey" json:"hunyuanAuthKey"`
+	// @Title zh-CN hunyuan api id for authorization
+	// @Description zh-CN 仅适用于Hun Yuan AI服务鉴权
+	hunyuanAuthId string `required:"false" yaml:"hunyuanAuthId" json:"hunyuanAuthId"`
+	// @Title zh-CN minimax group id
+	// @Description zh-CN 仅适用于minimax使用ChatCompletion Pro接口的模型
+	minimaxGroupId string `required:"false" yaml:"minimaxGroupId" json:"minimaxGroupId"`
 	// @Title zh-CN 模型名称映射表
 	// @Description zh-CN 用于将请求中的模型名称映射为目标AI服务商支持的模型名称。支持通过“*”来配置全局映射
 	modelMapping map[string]string `required:"false" yaml:"modelMapping" json:"modelMapping"`
@@ -128,6 +164,15 @@ type ProviderConfig struct {
 	// @Title zh-CN 模型对话上下文
 	// @Description zh-CN 配置一个外部获取对话上下文的文件来源，用于在AI请求中补充对话上下文
 	context *ContextConfig `required:"false" yaml:"context" json:"context"`
+	// @Title zh-CN 版本
+	// @Description zh-CN 请求AI服务的版本，目前仅适用于Claude AI服务
+	claudeVersion string `required:"false" yaml:"version" json:"version"`
+	// @Title zh-CN Cloudflare Account ID
+	// @Description zh-CN 仅适用于 Cloudflare Workers AI 服务。参考：https://developers.cloudflare.com/workers-ai/get-started/rest-api/#2-run-a-model-via-api
+	cloudflareAccountId string `required:"false" yaml:"cloudflareAccountId" json:"cloudflareAccountId"`
+	// @Title zh-CN Gemini AI内容过滤和安全级别设定
+	// @Description zh-CN 仅适用于 Gemini AI 服务。参考：https://ai.google.dev/gemini-api/docs/safety-settings
+	geminiSafetySetting map[string]string `required:"false" yaml:"geminiSafetySetting" json:"geminiSafetySetting"`
 }

 func (c *ProviderConfig) FromJson(json gjson.Result) {
@@ -140,6 +185,7 @@ func (c *ProviderConfig) FromJson(json gjson.Result) {
 	if c.timeout == 0 {
 		c.timeout = defaultTimeout
 	}
+	c.openaiCustomUrl = json.Get("openaiCustomUrl").String()
 	c.moonshotFileId = json.Get("moonshotFileId").String()
 	c.azureServiceUrl = json.Get("azureServiceUrl").String()
 	c.qwenFileIds = make([]string, 0)
@@ -162,12 +208,20 @@ func (c *ProviderConfig) FromJson(json gjson.Result) {
 		c.context = &ContextConfig{}
 		c.context.FromJson(contextJson)
 	}
+	c.claudeVersion = json.Get("claudeVersion").String()
+	c.hunyuanAuthId = json.Get("hunyuanAuthId").String()
+	c.hunyuanAuthKey = json.Get("hunyuanAuthKey").String()
+	c.minimaxGroupId = json.Get("minimaxGroupId").String()
+	c.cloudflareAccountId = json.Get("cloudflareAccountId").String()
+	if c.typ == providerTypeGemini {
+		c.geminiSafetySetting = make(map[string]string)
+		for k, v := range json.Get("geminiSafetySetting").Map() {
+			c.geminiSafetySetting[k] = v.String()
+		}
+	}
 }

 func (c *ProviderConfig) Validate() error {
-	if c.apiTokens == nil || len(c.apiTokens) == 0 {
-		return errors.New("no apiToken found in provider config")
-	}
 	if c.timeout < 0 {
 		return errors.New("invalid timeout in config")
 	}
@@ -182,7 +236,6 @@ func (c *ProviderConfig) Validate() error {

 	if c.typ == "" {
 		return errors.New("missing type in provider config")
-    
 	}
 	initializer, has := providerInitializers[c.typ]
 	if !has {
@@ -216,16 +269,38 @@ func CreateProvider(pc ProviderConfig) (Provider, error) {
 }

 func getMappedModel(model string, modelMapping map[string]string, log wrapper.Log) string {
-	if modelMapping == nil || len(modelMapping) == 0 {
-		return model
-	}
-	if v, ok := modelMapping[model]; ok && len(v) != 0 {
-		log.Debugf("model %s is mapped to %s explictly", model, v)
-		return v
-	}
-	if v, ok := modelMapping[wildcard]; ok {
-		log.Debugf("model %s is mapped to %s via wildcard", model, v)
-		return v
+	mappedModel := doGetMappedModel(model, modelMapping, log)
+	if len(mappedModel) != 0 {
+		return mappedModel
 	}
 	return model
 }
+
+func doGetMappedModel(model string, modelMapping map[string]string, log wrapper.Log) string {
+	if modelMapping == nil || len(modelMapping) == 0 {
+		return ""
+	}
+
+	if v, ok := modelMapping[model]; ok {
+		log.Debugf("model [%s] is mapped to [%s] explictly", model, v)
+		return v
+	}
+
+	for k, v := range modelMapping {
+		if k == wildcard || !strings.HasSuffix(k, wildcard) {
+			continue
+		}
+		k = strings.TrimSuffix(k, wildcard)
+		if strings.HasPrefix(model, k) {
+			log.Debugf("model [%s] is mapped to [%s] via prefix [%s]", model, v, k)
+			return v
+		}
+	}
+
+	if v, ok := modelMapping[wildcard]; ok {
+		log.Debugf("model [%s] is mapped to [%s] via wildcard", model, v)
+		return v
+	}
+
+	return ""
+}
--- a/plugins/wasm-go/extensions/ai-proxy/provider/qwen.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/qwen.go
@@ -5,6 +5,7 @@ import (
 	"errors"
 	"fmt"
 	"math"
+	"reflect"
 	"strings"
 	"time"

@@ -21,6 +22,7 @@ const (

 	qwenDomain             = "dashscope.aliyuncs.com"
 	qwenChatCompletionPath = "/api/v1/services/aigc/text-generation/generation"
+	qwenTextEmbeddingPath  = "/api/v1/services/embeddings/text-embedding/text-embedding"

 	qwenTopPMin = 0.000001
 	qwenTopPMax = 0.999999
@@ -37,6 +39,9 @@ func (m *qwenProviderInitializer) ValidateConfig(config ProviderConfig) error {
 	if len(config.qwenFileIds) != 0 && config.context != nil {
 		return errors.New("qwenFileIds and context cannot be configured at the same time")
 	}
+	if config.apiTokens == nil || len(config.apiTokens) == 0 {
+		return errors.New("no apiToken found in provider config")
+	}
 	return nil
 }

@@ -58,15 +63,17 @@ func (m *qwenProvider) GetProviderType() string {
 }

 func (m *qwenProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
-	if apiName != ApiNameChatCompletion {
+	if apiName == ApiNameChatCompletion {
+		_ = util.OverwriteRequestPath(qwenChatCompletionPath)
+	} else if apiName == ApiNameEmbeddings {
+		_ = util.OverwriteRequestPath(qwenTextEmbeddingPath)
+	} else {
 		return types.ActionContinue, errUnsupportedApiName
 	}
-	_ = util.OverwriteRequestPath(qwenChatCompletionPath)
 	_ = util.OverwriteRequestHost(qwenDomain)
-	_ = proxywasm.ReplaceHttpRequestHeader("Authorization", "Bearer "+m.config.GetRandomToken())
+	_ = util.OverwriteRequestAuthorization("Bearer " + m.config.GetRandomToken())

-	if m.config.protocol == protocolOriginal && m.config.context == nil {
-		ctx.DontReadRequestBody()
+	if m.config.protocol == protocolOriginal {
 		return types.ActionContinue, nil
 	}

@@ -78,10 +85,16 @@ func (m *qwenProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName ApiName
 }

 func (m *qwenProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
-	if apiName != ApiNameChatCompletion {
-		return types.ActionContinue, errUnsupportedApiName
+	if apiName == ApiNameChatCompletion {
+		return m.onChatCompletionRequestBody(ctx, body, log)
 	}
+	if apiName == ApiNameEmbeddings {
+		return m.onEmbeddingsRequestBody(ctx, body, log)
+	}
+	return types.ActionContinue, errUnsupportedApiName
+}

+func (m *qwenProvider) onChatCompletionRequestBody(ctx wrapper.HttpContext, body []byte, log wrapper.Log) (types.Action, error) {
 	if m.config.protocol == protocolOriginal {
 		if m.config.context == nil {
 			return types.ActionContinue, nil
@@ -99,11 +112,11 @@ func (m *qwenProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, b

 			if err != nil {
 				log.Errorf("failed to load context file: %v", err)
-				_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+				_ = util.SendResponse(500, "ai-proxy.qwen.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
 			}
 			m.insertContextMessage(request, content, false)
 			if err := replaceJsonRequestBody(request, log); err != nil {
-				_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+				_ = util.SendResponse(500, "ai-proxy.qwen.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
 			}
 		}, log)
 		if err == nil {
@@ -152,7 +165,7 @@ func (m *qwenProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, b
 		}()
 		if err != nil {
 			log.Errorf("failed to load context file: %v", err)
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.qwen.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
 		}
 		insertContextMessage(request, content)
 		qwenRequest := m.buildQwenTextGenerationRequest(request, streaming)
@@ -160,7 +173,7 @@ func (m *qwenProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, b
 			ctx.SetContext(ctxKeyIncrementalStreaming, qwenRequest.Parameters.IncrementalOutput)
 		}
 		if err := replaceJsonRequestBody(qwenRequest, log); err != nil {
-			_ = util.SendResponse(500, util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+			_ = util.SendResponse(500, "ai-proxy.qwen.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
 		}
 	}, log)
 	if err == nil {
@@ -169,6 +182,33 @@ func (m *qwenProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, b
 	return types.ActionContinue, err
 }

+func (m *qwenProvider) onEmbeddingsRequestBody(ctx wrapper.HttpContext, body []byte, log wrapper.Log) (types.Action, error) {
+	request := &embeddingsRequest{}
+	if err := json.Unmarshal(body, request); err != nil {
+		return types.ActionContinue, fmt.Errorf("unable to unmarshal request: %v", err)
+	}
+
+	log.Debugf("=== embeddings request: %v", request)
+
+	model := request.Model
+	if model == "" {
+		return types.ActionContinue, errors.New("missing model in the request")
+	}
+	ctx.SetContext(ctxKeyOriginalRequestModel, model)
+	mappedModel := getMappedModel(model, m.config.modelMapping, log)
+	if mappedModel == "" {
+		return types.ActionContinue, errors.New("model becomes empty after applying the configured mapping")
+	}
+	request.Model = mappedModel
+	ctx.SetContext(ctxKeyFinalRequestModel, request.Model)
+
+	if qwenRequest, err := m.buildQwenTextEmbeddingRequest(request); err == nil {
+		return types.ActionContinue, replaceJsonRequestBody(qwenRequest, log)
+	} else {
+		return types.ActionContinue, err
+	}
+}
+
 func (m *qwenProvider) OnResponseHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
 	if m.config.protocol == protocolOriginal {
 		ctx.DontReadResponseBody()
@@ -180,15 +220,16 @@ func (m *qwenProvider) OnResponseHeaders(ctx wrapper.HttpContext, apiName ApiNam
 }

 func (m *qwenProvider) OnStreamingResponseBody(ctx wrapper.HttpContext, name ApiName, chunk []byte, isLastChunk bool, log wrapper.Log) ([]byte, error) {
+	if name != ApiNameChatCompletion {
+		return chunk, nil
+	}
+
 	receivedBody := chunk
 	if bufferedStreamingBody, has := ctx.GetContext(ctxKeyStreamingBody).([]byte); has {
 		receivedBody = append(bufferedStreamingBody, chunk...)
 	}

-	incrementalStreaming, err := ctx.GetContext(ctxKeyIncrementalStreaming).(bool)
-	if !err {
-		incrementalStreaming = false
-	}
+	incrementalStreaming := ctx.GetBoolContext(ctxKeyIncrementalStreaming, false)

 	eventStartIndex, lineStartIndex, valueStartIndex := -1, -1, -1

@@ -264,6 +305,16 @@ func (m *qwenProvider) OnStreamingResponseBody(ctx wrapper.HttpContext, name Api
 }

 func (m *qwenProvider) OnResponseBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	if apiName == ApiNameChatCompletion {
+		return m.onChatCompletionResponseBody(ctx, body, log)
+	}
+	if apiName == ApiNameEmbeddings {
+		return m.onEmbeddingsResponseBody(ctx, body, log)
+	}
+	return types.ActionContinue, errUnsupportedApiName
+}
+
+func (m *qwenProvider) onChatCompletionResponseBody(ctx wrapper.HttpContext, body []byte, log wrapper.Log) (types.Action, error) {
 	qwenResponse := &qwenTextGenResponse{}
 	if err := json.Unmarshal(body, qwenResponse); err != nil {
 		return types.ActionContinue, fmt.Errorf("unable to unmarshal Qwen response: %v", err)
@@ -272,6 +323,15 @@ func (m *qwenProvider) OnResponseBody(ctx wrapper.HttpContext, apiName ApiName,
 	return types.ActionContinue, replaceJsonResponseBody(response, log)
 }

+func (m *qwenProvider) onEmbeddingsResponseBody(ctx wrapper.HttpContext, body []byte, log wrapper.Log) (types.Action, error) {
+	qwenResponse := &qwenTextEmbeddingResponse{}
+	if err := json.Unmarshal(body, qwenResponse); err != nil {
+		return types.ActionContinue, fmt.Errorf("unable to unmarshal Qwen response: %v", err)
+	}
+	response := m.buildEmbeddingsResponse(ctx, qwenResponse)
+	return types.ActionContinue, replaceJsonResponseBody(response, log)
+}
+
 func (m *qwenProvider) buildQwenTextGenerationRequest(origRequest *chatCompletionRequest, streaming bool) *qwenTextGenRequest {
 	messages := make([]qwenMessage, 0, len(origRequest.Messages))
 	for i := range origRequest.Messages {
@@ -324,11 +384,11 @@ func (m *qwenProvider) buildChatCompletionResponse(ctx wrapper.HttpContext, qwen
 	return &chatCompletionResponse{
 		Id:                qwenResponse.RequestId,
 		Created:           time.Now().UnixMilli() / 1000,
-		Model:             ctx.GetContext(ctxKeyFinalRequestModel).(string),
+		Model:             ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
 		SystemFingerprint: "",
 		Object:            objectChatCompletion,
 		Choices:           choices,
-		Usage: chatCompletionUsage{
+		Usage: usage{
 			PromptTokens:     qwenResponse.Usage.InputTokens,
 			CompletionTokens: qwenResponse.Usage.OutputTokens,
 			TotalTokens:      qwenResponse.Usage.TotalTokens,
@@ -340,7 +400,7 @@ func (m *qwenProvider) buildChatCompletionStreamingResponse(ctx wrapper.HttpCont
 	baseMessage := chatCompletionResponse{
 		Id:                qwenResponse.RequestId,
 		Created:           time.Now().UnixMilli() / 1000,
-		Model:             ctx.GetContext(ctxKeyFinalRequestModel).(string),
+		Model:             ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
 		Choices:           make([]chatCompletionChoice, 0),
 		SystemFingerprint: "",
 		Object:            objectChatCompletionChunk,
@@ -349,11 +409,19 @@ func (m *qwenProvider) buildChatCompletionStreamingResponse(ctx wrapper.HttpCont
 	responses := make([]*chatCompletionResponse, 0)

 	qwenChoice := qwenResponse.Output.Choices[0]
+	// Yes, Qwen uses a string "null" as null.
+	finished := qwenChoice.FinishReason != "" && qwenChoice.FinishReason != "null"
 	message := qwenChoice.Message

 	deltaContentMessage := &chatMessage{Role: message.Role, Content: message.Content}
 	deltaToolCallsMessage := &chatMessage{Role: message.Role, ToolCalls: append([]toolCall{}, message.ToolCalls...)}
 	if !incrementalStreaming {
+		for _, tc := range message.ToolCalls {
+			if tc.Function.Arguments == "" && !finished {
+				// We don't push any tool call until its arguments are available.
+				return nil
+			}
+		}
 		if pushedMessage, ok := ctx.GetContext(ctxKeyPushedMessage).(qwenMessage); ok {
 			if message.Content == "" {
 				message.Content = pushedMessage.Content
@@ -386,13 +454,13 @@ func (m *qwenProvider) buildChatCompletionStreamingResponse(ctx wrapper.HttpCont
 		responses = append(responses, &response)
 	}

-	// Yes, Qwen uses a string "null" as null.
-	if qwenChoice.FinishReason != "" && qwenChoice.FinishReason != "null" {
+	if finished {
 		finishResponse := *&baseMessage
-		finishResponse.Choices = append(finishResponse.Choices, chatCompletionChoice{FinishReason: qwenChoice.FinishReason})
+		finishResponse.Choices = append(finishResponse.Choices, chatCompletionChoice{Delta: &chatMessage{}, FinishReason: qwenChoice.FinishReason})

 		usageResponse := *&baseMessage
-		usageResponse.Usage = chatCompletionUsage{
+		usageResponse.Choices = []chatCompletionChoice{{Delta: &chatMessage{}}}
+		usageResponse.Usage = usage{
 			PromptTokens:     qwenResponse.Usage.InputTokens,
 			CompletionTokens: qwenResponse.Usage.OutputTokens,
 			TotalTokens:      qwenResponse.Usage.TotalTokens,
@@ -477,6 +545,50 @@ func (m *qwenProvider) appendStreamEvent(responseBuilder *strings.Builder, event
 	responseBuilder.WriteString("\n\n")
 }

+func (m *qwenProvider) buildQwenTextEmbeddingRequest(request *embeddingsRequest) (*qwenTextEmbeddingRequest, error) {
+	var texts []string
+	if str, isString := request.Input.(string); isString {
+		texts = []string{str}
+	} else if strs, isArray := request.Input.([]interface{}); isArray {
+		texts = make([]string, 0, len(strs))
+		for _, item := range strs {
+			if str, isString := item.(string); isString {
+				texts = append(texts, str)
+			} else {
+				return nil, errors.New("unsupported input type in array: " + reflect.TypeOf(item).String())
+			}
+		}
+	} else {
+		return nil, errors.New("unsupported input type: " + reflect.TypeOf(request.Input).String())
+	}
+	return &qwenTextEmbeddingRequest{
+		Model: request.Model,
+		Input: qwenTextEmbeddingInput{
+			Texts: texts,
+		},
+	}, nil
+}
+
+func (m *qwenProvider) buildEmbeddingsResponse(ctx wrapper.HttpContext, qwenResponse *qwenTextEmbeddingResponse) *embeddingsResponse {
+	data := make([]embedding, 0, len(qwenResponse.Output.Embeddings))
+	for _, qwenEmbedding := range qwenResponse.Output.Embeddings {
+		data = append(data, embedding{
+			Object:    "embedding",
+			Index:     qwenEmbedding.TextIndex,
+			Embedding: qwenEmbedding.Embedding,
+		})
+	}
+	return &embeddingsResponse{
+		Object: "list",
+		Data:   data,
+		Model:  ctx.GetContext(ctxKeyFinalRequestModel).(string),
+		Usage: usage{
+			PromptTokens: qwenResponse.Usage.TotalTokens,
+			TotalTokens:  qwenResponse.Usage.TotalTokens,
+		},
+	}
+}
+
 type qwenTextGenRequest struct {
 	Model      string                `json:"model"`
 	Input      qwenTextGenInput      `json:"input"`
@@ -503,7 +615,7 @@ type qwenTextGenParameters struct {
 type qwenTextGenResponse struct {
 	RequestId string            `json:"request_id"`
 	Output    qwenTextGenOutput `json:"output"`
-	Usage     qwenTextGenUsage  `json:"usage"`
+	Usage     qwenUsage         `json:"usage"`
 }

 type qwenTextGenOutput struct {
@@ -516,7 +628,7 @@ type qwenTextGenChoice struct {
 	Message      qwenMessage `json:"message"`
 }

-type qwenTextGenUsage struct {
+type qwenUsage struct {
 	InputTokens  int `json:"input_tokens"`
 	OutputTokens int `json:"output_tokens"`
 	TotalTokens  int `json:"total_tokens"`
@@ -529,6 +641,36 @@ type qwenMessage struct {
 	ToolCalls []toolCall `json:"tool_calls,omitempty"`
 }

+type qwenTextEmbeddingRequest struct {
+	Model      string                      `json:"model"`
+	Input      qwenTextEmbeddingInput      `json:"input"`
+	Parameters qwenTextEmbeddingParameters `json:"parameters,omitempty"`
+}
+
+type qwenTextEmbeddingInput struct {
+	Texts []string `json:"texts"`
+}
+
+type qwenTextEmbeddingParameters struct {
+	TextType string `json:"text_type,omitempty"`
+}
+
+type qwenTextEmbeddingResponse struct {
+	RequestId string                  `json:"request_id"`
+	Output    qwenTextEmbeddingOutput `json:"output"`
+	Usage     qwenUsage               `json:"usage"`
+}
+
+type qwenTextEmbeddingOutput struct {
+	RequestId  string               `json:"request_id"`
+	Embeddings []qwenTextEmbeddings `json:"embeddings"`
+}
+
+type qwenTextEmbeddings struct {
+	TextIndex int       `json:"text_index"`
+	Embedding []float64 `json:"embedding"`
+}
+
 func qwenMessageToChatMessage(qwenMessage qwenMessage) chatMessage {
 	return chatMessage{
 		Name:      qwenMessage.Name,
--- a/plugins/wasm-go/extensions/ai-proxy/provider/request_helper.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/request_helper.go
@@ -2,7 +2,6 @@ package provider

 import (
 	"encoding/json"
-	"errors"
 	"fmt"

 	"github.com/alibaba/higress/plugins/wasm-go/pkg/wrapper"
@@ -14,7 +13,7 @@ func decodeChatCompletionRequest(body []byte, request *chatCompletionRequest) er
 		return fmt.Errorf("unable to unmarshal request: %v", err)
 	}
 	if request.Messages == nil || len(request.Messages) == 0 {
-		return errors.New("no message found in the request body")
+		return fmt.Errorf("no message found in the request body: %s", body)
 	}
 	return nil
 }
--- a/plugins/wasm-go/extensions/ai-proxy/provider/spark.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/spark.go
@@ -0,0 +1,207 @@
+package provider
+
+import (
+	"encoding/json"
+	"errors"
+	"fmt"
+	"strings"
+	"time"
+
+	"github.com/alibaba/higress/plugins/wasm-go/extensions/ai-proxy/util"
+	"github.com/alibaba/higress/plugins/wasm-go/pkg/wrapper"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm/types"
+)
+
+// sparkProvider is the provider for SparkLLM AI service.
+const (
+	sparkHost               = "spark-api-open.xf-yun.com"
+	sparkChatCompletionPath = "/v1/chat/completions"
+)
+
+type sparkProviderInitializer struct {
+}
+
+type sparkProvider struct {
+	config       ProviderConfig
+	contextCache *contextCache
+}
+
+type sparkRequest struct {
+	Model       string        `json:"model"`
+	Messages    []chatMessage `json:"messages"`
+	MaxTokens   int           `json:"max_tokens,omitempty"`
+	TopK        int           `json:"top_k,omitempty"`
+	Stream      bool          `json:"stream,omitempty"`
+	Temperature float64       `json:"temperature,omitempty"`
+	Tools       []tool        `json:"tools,omitempty"`
+	ToolChoice  string        `json:"tool_choice,omitempty"`
+}
+
+type sparkResponse struct {
+	Code    int                    `json:"code"`
+	Message string                 `json:"message"`
+	Sid     string                 `json:"sid"`
+	Choices []chatCompletionChoice `json:"choices"`
+	Usage   usage                  `json:"usage,omitempty"`
+}
+
+type sparkStreamResponse struct {
+	sparkResponse
+	Id      string `json:"id"`
+	Created int64  `json:"created"`
+}
+
+func (i *sparkProviderInitializer) ValidateConfig(config ProviderConfig) error {
+	return nil
+}
+
+func (i *sparkProviderInitializer) CreateProvider(config ProviderConfig) (Provider, error) {
+	return &sparkProvider{
+		config:       config,
+		contextCache: createContextCache(&config),
+	}, nil
+}
+
+func (p *sparkProvider) GetProviderType() string {
+	return providerTypeSpark
+}
+
+func (p *sparkProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+	_ = util.OverwriteRequestHost(sparkHost)
+	_ = util.OverwriteRequestPath(sparkChatCompletionPath)
+	_ = util.OverwriteRequestAuthorization("Bearer " + p.config.GetRandomToken())
+	_ = proxywasm.RemoveHttpRequestHeader("Accept-Encoding")
+	_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
+	return types.ActionContinue, nil
+}
+
+func (p *sparkProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+	// 使用Spark协议
+	if p.config.protocol == protocolOriginal {
+		request := &sparkRequest{}
+		if err := json.Unmarshal(body, request); err != nil {
+			return types.ActionContinue, fmt.Errorf("unable to unmarshal request: %v", err)
+		}
+		if request.Model == "" {
+			return types.ActionContinue, errors.New("request model is empty")
+		}
+		// 目前星火在模型名称错误时，也会调用generalv3，这里还是按照输入的模型名称设置响应里的模型名称
+		ctx.SetContext(ctxKeyFinalRequestModel, request.Model)
+		return types.ActionContinue, replaceJsonRequestBody(request, log)
+	} else {
+		// 使用openai协议
+		request := &chatCompletionRequest{}
+		if err := decodeChatCompletionRequest(body, request); err != nil {
+			return types.ActionContinue, err
+		}
+		if request.Model == "" {
+			return types.ActionContinue, errors.New("missing model in chat completion request")
+		}
+		// 映射模型
+		mappedModel := getMappedModel(request.Model, p.config.modelMapping, log)
+		if mappedModel == "" {
+			return types.ActionContinue, errors.New("model becomes empty after applying the configured mapping")
+		}
+		ctx.SetContext(ctxKeyFinalRequestModel, mappedModel)
+		request.Model = mappedModel
+		return types.ActionContinue, replaceJsonRequestBody(request, log)
+	}
+}
+
+func (p *sparkProvider) OnResponseHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	_ = proxywasm.RemoveHttpResponseHeader("Content-Length")
+	return types.ActionContinue, nil
+}
+
+func (p *sparkProvider) OnResponseBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	sparkResponse := &sparkResponse{}
+	if err := json.Unmarshal(body, sparkResponse); err != nil {
+		return types.ActionContinue, fmt.Errorf("unable to unmarshal spark response: %v", err)
+	}
+	if sparkResponse.Code != 0 {
+		return types.ActionContinue, fmt.Errorf("spark response error, error_code: %d, error_message: %s", sparkResponse.Code, sparkResponse.Message)
+	}
+	response := p.responseSpark2OpenAI(ctx, sparkResponse)
+	return types.ActionContinue, replaceJsonResponseBody(response, log)
+}
+
+func (p *sparkProvider) OnStreamingResponseBody(ctx wrapper.HttpContext, name ApiName, chunk []byte, isLastChunk bool, log wrapper.Log) ([]byte, error) {
+	if isLastChunk || len(chunk) == 0 {
+		return nil, nil
+	}
+	responseBuilder := &strings.Builder{}
+	lines := strings.Split(string(chunk), "\n")
+	for _, data := range lines {
+		if len(data) < 6 {
+			// ignore blank line or wrong format
+			continue
+		}
+		data = data[6:]
+		// The final response is `data: [DONE]`
+		if data == "[DONE]" {
+			continue
+		}
+		var sparkResponse sparkStreamResponse
+		if err := json.Unmarshal([]byte(data), &sparkResponse); err != nil {
+			log.Errorf("unable to unmarshal spark response: %v", err)
+			continue
+		}
+		response := p.streamResponseSpark2OpenAI(ctx, &sparkResponse)
+		responseBody, err := json.Marshal(response)
+		if err != nil {
+			log.Errorf("unable to marshal response: %v", err)
+			return nil, err
+		}
+		p.appendResponse(responseBuilder, string(responseBody))
+	}
+	modifiedResponseChunk := responseBuilder.String()
+	log.Debugf("=== modified response chunk: %s", modifiedResponseChunk)
+	return []byte(modifiedResponseChunk), nil
+}
+
+func (p *sparkProvider) responseSpark2OpenAI(ctx wrapper.HttpContext, response *sparkResponse) *chatCompletionResponse {
+	choices := make([]chatCompletionChoice, len(response.Choices))
+	for idx, c := range response.Choices {
+		choices[idx] = chatCompletionChoice{
+			Index:   c.Index,
+			Message: &chatMessage{Role: c.Message.Role, Content: c.Message.Content},
+		}
+	}
+	return &chatCompletionResponse{
+		Id:      response.Sid,
+		Created: time.Now().UnixMilli() / 1000,
+		Object:  objectChatCompletion,
+		Model:   ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
+		Choices: choices,
+		Usage:   response.Usage,
+	}
+}
+
+func (p *sparkProvider) streamResponseSpark2OpenAI(ctx wrapper.HttpContext, response *sparkStreamResponse) *chatCompletionResponse {
+	choices := make([]chatCompletionChoice, len(response.Choices))
+	for idx, c := range response.Choices {
+		choices[idx] = chatCompletionChoice{
+			Index: c.Index,
+			Delta: &chatMessage{Role: c.Delta.Role, Content: c.Delta.Content},
+		}
+	}
+	return &chatCompletionResponse{
+		Id:      response.Sid,
+		Created: response.Created,
+		Model:   ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
+		Object:  objectChatCompletion,
+		Choices: choices,
+		Usage:   response.Usage,
+	}
+}
+
+func (p *sparkProvider) appendResponse(responseBuilder *strings.Builder, responseBody string) {
+	responseBuilder.WriteString(fmt.Sprintf("%s %s\n\n", streamDataItemKey, responseBody))
+}
--- a/plugins/wasm-go/extensions/ai-proxy/provider/stepfun.go
+++ b/plugins/wasm-go/extensions/ai-proxy/provider/stepfun.go
@@ -0,0 +1,83 @@
+package provider
+
+import (
+	"errors"
+	"fmt"
+
+	"github.com/alibaba/higress/plugins/wasm-go/extensions/ai-proxy/util"
+	"github.com/alibaba/higress/plugins/wasm-go/pkg/wrapper"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm"
+	"github.com/higress-group/proxy-wasm-go-sdk/proxywasm/types"
+)
+
+const (
+	stepfunDomain             = "api.stepfun.com"
+	stepfunChatCompletionPath = "/v1/chat/completions"
+)
+
+type stepfunProviderInitializer struct {
+}
+
+func (m *stepfunProviderInitializer) ValidateConfig(config ProviderConfig) error {
+	if config.apiTokens == nil || len(config.apiTokens) == 0 {
+		return errors.New("no apiToken found in provider config")
+	}
+	return nil
+}
+
+func (m *stepfunProviderInitializer) CreateProvider(config ProviderConfig) (Provider, error) {
+	return &stepfunProvider{
+		config:       config,
+		contextCache: createContextCache(&config),
+	}, nil
+}
+
+type stepfunProvider struct {
+	config       ProviderConfig
+	contextCache *contextCache
+}
+
+func (m *stepfunProvider) GetProviderType() string {
+	return providerTypeStepfun
+}
+
+func (m *stepfunProvider) OnRequestHeaders(ctx wrapper.HttpContext, apiName ApiName, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+	_ = util.OverwriteRequestPath(stepfunChatCompletionPath)
+	_ = util.OverwriteRequestHost(stepfunDomain)
+	_ = util.OverwriteRequestAuthorization("Bearer " + m.config.GetRandomToken())
+	_ = proxywasm.RemoveHttpRequestHeader("Content-Length")
+	return types.ActionContinue, nil
+}
+
+func (m *stepfunProvider) OnRequestBody(ctx wrapper.HttpContext, apiName ApiName, body []byte, log wrapper.Log) (types.Action, error) {
+	if apiName != ApiNameChatCompletion {
+		return types.ActionContinue, errUnsupportedApiName
+	}
+	if m.contextCache == nil {
+		return types.ActionContinue, nil
+	}
+	request := &chatCompletionRequest{}
+	if err := decodeChatCompletionRequest(body, request); err != nil {
+		return types.ActionContinue, err
+	}
+	err := m.contextCache.GetContent(func(content string, err error) {
+		defer func() {
+			_ = proxywasm.ResumeHttpRequest()
+		}()
+		if err != nil {
+			log.Errorf("failed to load context file: %v", err)
+			_ = util.SendResponse(500, "ai-proxy.stepfun.load_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to load context file: %v", err))
+		}
+		insertContextMessage(request, content)
+		if err := replaceJsonRequestBody(request, log); err != nil {
+			_ = util.SendResponse(500, "ai-proxy.stepfun.insert_ctx_failed", util.MimeTypeTextPlain, fmt.Sprintf("failed to replace request body: %v", err))
+		}
+	}, log)
+	if err == nil {
+		return types.ActionPause, nil
+	}
+	return types.ActionContinue, err
+}
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
韩贤涛	04a9104062	feat: support gemini ai model (#1173 )	2024-08-09 09:55:40 +08:00
Se7en	564f8c770a	fix: fix tracing configmap template to handle initial installation (#1191 )	2024-08-09 08:29:51 +08:00
Se7en	fec2e9dfc9	feat: improve Skywalking and Zipkin integration (#1131 )	2024-08-08 22:40:33 +08:00
Jingze	dc4ddb52ee	fix bug of empty config plugin still start (#1189 )	2024-08-08 18:04:47 +08:00
Jun	6f221ead53	feat:add service rule match for wasmplugin in control panel (#1166 )	2024-08-08 18:04:33 +08:00
韩贤涛	53f8410843	feat: ext auth forward_auth endpoint_mode enhancement (#1180 )	2024-08-08 18:01:51 +08:00
rinfx	a17ac9e4c6	Optimize ai-rag plugin (#1170 )	2024-08-08 18:00:02 +08:00
澄潭	5e95f6f057	Update README.md	2024-08-08 17:14:18 +08:00
澄潭	94f29e56c0	Update README.md	2024-08-08 17:12:33 +08:00
澄潭	870157c576	Update README.md	2024-08-08 15:53:21 +08:00
urlyy	c78ef7011d	Feat: Add Spark llm support for plugins/ai-proxy (#1139 )	2024-08-08 15:16:58 +08:00
澄潭	dc0dcaaaee	azure-openai support other type api (#1187 )	2024-08-08 13:33:12 +08:00
EricaLiu	34f5722d93	fix: add support for nacos triple protocol (#1186 )	2024-08-08 10:29:48 +08:00
澄潭	55fdddee2f	optimize transformer plugin (#1183 )	2024-08-08 09:46:11 +08:00
007gzs	980ffde244	Optimize WASM Rust SDK's body caching logic. (#1181 )	2024-08-07 20:06:11 +08:00
澄潭	0a578c2a04	ai-proxy: support custom openai provider (#1176 ) Co-authored-by: Kent Dong <ch3cho@qq.com>	2024-08-07 10:33:01 +08:00
澄潭	536a3069a8	Update README.md	2024-08-06 20:15:33 +08:00
韩贤涛	08c64ed467	fix:fix bug in ext-auth wasm plugin (#1152 )	2024-08-05 11:04:31 +08:00
澄潭	cc74c0da93	replace regexp (#1169 )	2024-07-31 17:48:38 +08:00
Kent Dong	210b97b06b	fix: Use the official tinygo package to build Wasm go plugin builder (#1161 )	2024-07-29 16:05:23 +08:00
007gzs	bccfbde621	fix PluginHttpWrapper 中 Context的回调未代理 . request-block case_sensitive 逻辑错误 (#1146 )	2024-07-27 10:25:14 +08:00
澄潭	f1c6e78047	Update Makefile.core.mk	2024-07-26 14:06:38 +08:00
澄潭	1c415c60c3	rel: Release v1.4.2 (#1159 )	2024-07-26 13:55:03 +08:00
澄潭	59acb61926	Update Makefile.core.mk	2024-07-26 13:50:16 +08:00
澄潭	29079f4e2a	Support set buffer limit (#1153 )	2024-07-25 20:42:39 +08:00
澄潭	95edce024d	support custom trace span tag (#1156 )	2024-07-25 20:42:09 +08:00
Kent Dong	b6d07a157c	feat: Always buffer request body in ai-proxy plugin (#1155 )	2024-07-25 19:35:39 +08:00
澄潭	10569f49ae	support keep original auth header (#1151 )	2024-07-24 19:31:38 +08:00
Jun	2a588c99c7	fix: add full push when higress-https configmap updated and fix certmagic storage (#1105 )	2024-07-24 19:30:40 +08:00
Kent Dong	0cfef34bff	feat: Support fallback route in ai-proxy plugin (#1123 )	2024-07-24 19:25:32 +08:00
rinfx	5c2b5d5750	potential bug fix (#1141 )	2024-07-24 19:24:47 +08:00
rinfx	8f483518a9	support take effect on api level (#1150 )	2024-07-24 19:23:55 +08:00
Kent Dong	f6ee4ed166	fix: Bypass the response body processing in ai-proxy if it is returned internally (#1149 )	2024-07-24 16:18:00 +08:00
Kent Dong	9a9e924037	feat: Make higress-core and higress-gateway as the default container (#1144 )	2024-07-24 11:25:16 +08:00
jiaomh	e7d66f691f	chore: Update multiple dependencies to the latest version (#1143 )	2024-07-23 11:48:44 +08:00
rinfx	8c48fcb423	update template decorator (#1142 )	2024-07-22 17:04:55 +08:00
007gzs	ef31e09310	feat: add rust demo plugin request block (#1091 ) Co-authored-by: Yi <lynskylate@gmail.com>	2024-07-22 15:49:06 +08:00
韩贤涛	c0f2cafdc8	feat: support ext_auth wasmplugin (#1103 )	2024-07-17 15:30:32 +08:00
Kent Dong	d5a9ff3a98	fix: Fix possible type-casting related panics in ai-proxy plugin (#1127 )	2024-07-16 18:38:43 +08:00
Kent Dong	f069ad5b0d	feat: Add statusCodeDetails info when returning response in Wasm plugins directly (#1116 )	2024-07-16 09:52:46 +08:00
xu.zhao	85219b6c53	fix: controller has no right to watch deployment (#1089 ) Co-authored-by: Kent Dong <ch3cho@qq.com>	2024-07-15 16:34:24 +08:00
mamba	5041277be3	feat: 🎸 add frontend gray plugin (#1120 ) Co-authored-by: Kent Dong <ch3cho@qq.com>	2024-07-15 15:47:04 +08:00
rinfx	c00c8827f9	support service-level match config (#1112 )	2024-07-15 14:00:02 +08:00
rinfx	46218058d1	token-ratelimit crash bugfix (#1119 )	2024-07-12 15:05:15 +08:00
Kent Dong	5306385e6b	feat: Support loading custom parameters in build-and-push-wasm-plugin-image.yaml (#1118 )	2024-07-12 14:23:12 +08:00
Se7en	4e881fdd3f	doc: update cluster-key-rate-limit doc (#1113 ) Co-authored-by: 澄潭 <zty98751@alibaba-inc.com> Co-authored-by: Kent Dong <ch3cho@qq.com>	2024-07-11 17:31:14 +08:00
Kent Dong	59aa3b5488	fix: Use "controller.name" to refer the controller service in higress-config (#1108 )	2024-07-11 16:14:16 +08:00
Kent Dong	c40cf85aad	fix: Fix the incorrect image name used in build-and-push-wasm-plugin-image.yaml (#1109 )	2024-07-10 13:51:13 +08:00
Kent Dong	7c749b864c	fix: Fix some bugs in build-and-push-wasm-plugin-image.yaml (#1107 )	2024-07-10 13:41:58 +08:00
Yiiong	74ddbf02f6	feat:add build-and-push-wasm-plugin-image.yaml (#1069 ) Co-authored-by: Kent Dong <ch3cho@qq.com>	2024-07-08 21:44:58 +08:00
zzjin	60c56a16ab	Support `CredentialConfig.TLSSecret` with namespace. Resolve: #1066 (#1095 ) Signed-off-by: zzjin <tczzjin@gmail.com>	2024-07-08 19:49:51 +08:00
Kent Dong	5a2c6835f7	feat: Support embeddings API for Qwen in the ai-proxy plugin (#1079 )	2024-07-08 19:37:08 +08:00
Kent Dong	12a5612450	feat: Support model prefix mapping in ai-proxy (#1097 )	2024-07-08 19:33:08 +08:00
nohup	b9f5c4d1f2	feat: support Cloudflare Workers AI (#1068 ) Co-authored-by: Kent Dong <ch3cho@qq.com>	2024-07-08 19:27:11 +08:00
Jun	d7bdcbd026	fix priorityClassName missed (#1096 )	2024-07-08 19:26:08 +08:00
野生程序员	dd284d1f24	feat: loadBalancerClass (#1071 )	2024-07-08 10:58:33 +08:00
jiaomh	a7ee523c98	Update test/README.md (#1098 )	2024-07-07 10:06:32 +08:00
Kent Dong	4bbfb131ee	feat: Load 3rd-party images from higress image repo (#1067 )	2024-07-04 20:14:00 +08:00
Se7en	6fd71f9749	fix: prometheus port (#1076 )	2024-07-03 13:46:32 +08:00
pepesi	e0159f501a	fix jwt-auth plugin claims_to_headers failed (#1075 )	2024-07-03 10:11:17 +08:00
Kent Dong	56226d5052	feat: Create an IngressClass resource in the helm chart (#1072 )	2024-07-02 21:22:00 +08:00
pepesi	086a9cc973	fixed ai-statistics plugin statistics error (#1060 )	2024-07-02 20:35:12 +08:00
Tao Jikun	e389313aa3	feat: update doc for running Ingress API conformance tests (#1065 ) Co-authored-by: Kent Dong <ch3cho@qq.com>	2024-06-27 14:55:10 +08:00
澄潭	f64c601264	compatiable with openai sdk (#1061 ) Co-authored-by: Kent Dong <ch3cho@qq.com>	2024-06-27 09:30:52 +08:00
co63oc	9c6ea109f8	Fix typos (#1053 )	2024-06-26 19:47:39 +08:00
yy	4ca2d23404	feat: helm charts support installing gateway in daemonset mod. (#1054 )	2024-06-26 19:47:20 +08:00
co63oc	0ce52de59b	Fix typos (#1050 )	2024-06-22 16:22:03 +08:00
澄潭	81e459da01	Update Makefile.core.mk	2024-06-19 17:40:22 +08:00
澄潭	63539ca15c	rel:Release v1.4.1 (#1048 )	2024-06-19 17:12:51 +08:00
澄潭	1eea75f130	Update Makefile.core.mk	2024-06-19 17:00:06 +08:00
The Wind	d333656cc3	feat: support summary output for route/cluster/listener in hgctl gateway-config command (#995 ) (#996 )	2024-06-19 13:55:59 +08:00
Se7en	51dca7055a	feat: support claude ai model (#969 ) Signed-off-by: chengzw <chengzw258@163.com>	2024-06-19 13:53:21 +08:00
Chi Kai	ab1bc0a73a	feat: support stepfun model (#1012 ) Co-authored-by: Kent Dong <kentstl@163.com>	2024-06-19 13:51:02 +08:00
Kent Dong	ffee7dc5ea	fix: Accommodate the incomplete function name in the initial event from Qwen (#1045 )	2024-06-19 13:48:20 +08:00
rinfx	1ea87f0e7a	add plugin: ai-token-ratelimit (#1015 )	2024-06-19 13:46:59 +08:00
rinfx	7164653446	ai rag updates (#1046 )	2024-06-19 13:46:17 +08:00
澄潭	2a1a391054	Optimize the effectiveness speed of xds and add AI-related metric tags (#1047 )	2024-06-19 13:45:42 +08:00
rinfx	0785d4aac4	add plugin: ai-statistics (#1011 )	2024-06-18 17:53:04 +08:00
rinfx	4ca4bec2b5	add plugin: ai-transformer (#1035 )	2024-06-18 17:51:38 +08:00
rinfx	174350d3fb	add plugin: ai-rag (#1038 )	2024-06-17 15:37:00 +08:00
rinfx	0380cb03d3	add plugin: ai-prompt-template (#1019 )	2024-06-17 15:33:05 +08:00
rinfx	15d9f76ff9	add plugin: ai-prompt-decorator (#1021 )	2024-06-17 15:31:34 +08:00
rinfx	5f15017963	add plugin: ai-security-guard (#1034 )	2024-06-17 10:41:46 +08:00
韩贤涛	634de3f7f8	feat: cluster key rate limit enhancement (#1036 )	2024-06-17 10:37:03 +08:00
韩贤涛	12cc44b324	feat: cluster key rate limit (#1002 )	2024-06-12 14:51:46 +08:00
韩贤涛	d53c713561	feat: support minimax ai model (#1033 )	2024-06-09 21:01:31 +08:00
澄潭	5acc6f73b2	fix prometheus stats (#1031 )	2024-06-07 17:24:19 +08:00
韩贤涛	2db0b60a98	feat: support baidu ernie bot ai model (#1024 ) Co-authored-by: Kent Dong <ch3cho@qq.com>	2024-06-06 18:19:55 +08:00
nash5	c6e3db95e0	feature: add hunyuan llm support for plugins/ai-proxy (#1018 ) Co-authored-by: Kent Dong <ch3cho@qq.com>	2024-06-06 18:11:51 +08:00
Ink33	ed976c6d06	feat(plugin): implement golang version of plugin jwt-auth (#743 ) Signed-off-by: Ink33 <Ink33@smlk.org>	2024-06-06 10:22:51 +08:00