feat: Add AI-based bidding information tool MCP service (#2343)

This commit is contained in:
hourmoneys
2025-06-16 10:14:46 +08:00
committed by GitHub
parent 7040e4bd34
commit 2670ecbf8e
3 changed files with 185 additions and 0 deletions

View File

@@ -0,0 +1,61 @@
# Bid Tools MCP Server
## What is Junrun Bidding Digital Employee?
- Bidding is an important channel for enterprises to obtain projects and customers. Junrun Bidding Digital Employee can provide comprehensive and accurate bidding information, helping enterprises effectively improve their bidding capabilities.
## What are the problems with traditional bidding information acquisition channels?
- Obtaining bidding information through bidding platforms was a commonly used method in the past. However, there are many sources of bidding information, and it is difficult for general bidding platforms to cover all of them. There is a large amount of duplication in the bidding information from different platforms, making it difficult to deduplicate and filter.
- Traditional bidding information screening only uses keyword filtering, which is not very accurate. A large amount of bidding information still needs to be screened by experienced personnel, resulting in high labor costs.
## What can Junrun Bidding Digital Employee do?
- According to the subscribed bidding information, it pushes accurate bidding information filtered by AI via email every day. You can view the title, details, and download attachments of the bidding information.
- Based on AI capabilities, it analyzes bidding documents and automatically extracts key information, greatly improving the efficiency of bidding decision-making.
- Based on historical bidding information, it provides future bidding predictions to help enterprises plan in advance.
- Analyzes the historical bidding and winning situations of customers to support the formulation of bidding strategies.
- Analyzes the historical bidding and winning situations of competitors to obtain a list of target customers.
## Why choose Junrun Bidding Digital Employee?
### More comprehensive bidding data
- It cooperates with multiple source data enterprises, covering government bidding information, third - party bidding information, and enterprise self - owned website bidding information. In total, it covers more than 100,000 first - release information source stations, more than 290,000 collection channel addresses, and more than 10,000 new media official accounts (continuously updated). Tens of thousands of bidding and procurement information are updated daily, and the coverage rate of publicly available bidding and procurement information across the network can reach over 98% (the remaining 2% is due to new resources and the restructuring of existing resources).
### More accurate screening results
- Based on AI large - scale models, data training, and matching technology, the accuracy is significantly improved compared to traditional keyword search and rule - based judgment.
### Higher bidding decision - making efficiency
- Analysis of historical winning situations and one - click analysis of bidding documents to extract key information effectively assist in bidding decision - making.
## Data Source Coverage
| Site Type | Number of Sites (100,000+) | Number of Collection Source Addresses (290,000+) | Information Release Volume Proportion |
|------------------|----------|------------|----------------|
| Enterprise Bidding | - |- | 13.1183% |
| E - Procurement Platform | - | - | 33.0614% |
| Financial Banks | - | - | 0.0757% |
| Higher Education | - | - | 0.9645% |
| Education Bureau (Website) | - | - | 0.0173% |
| Hospitals | - | - | 0.8030% |
| Other Healthcare Institutions | - | - | 0.0167% |
| People's Government | - | - | 8.6457% |
| Public Resource Centers | - | - | 11.6248% |
| Government Procurement Centers | - | - | 27.8139% |
| Project Trading Centers | - | - | 3.5826% |
| Administrative Service Centers | - | - | 0.2761% |
## Features
- `getBidlist`: Query bidding information across the network based on keywords. Input multiple keywords and return the bidding query results.
- `getBidinfo`: Query the details of bidding information based on the bidding ID. Input the bidding ID and return the details of the bidding information.
- `Email`: An offline - configured enterprise bidding matching email notification service. It combines the enterprise's industry, business scope, keyword groups, etc. to create a unique enterprise profile. Then, it uses the Qwen3 large model of Alibaba's Bailian platform to deeply analyze daily bidding information and provide accurate bidding matching services for enterprises.
## Tutorial
### Get API Key
1. Register an account [Create an ID](https://moonai-bid.junrunrenli.com?src=higress)
2. Send an email to: yuanpeng@junrunrenli.com with the subject: MCP and the content: Apply for the bidding tool service, and provide your account.
### Knowledge Base (Alibaba Bailian Knowledge Base)
1. The clearer the enterprise's exclusive profile, the better the matching effect. Through continuous optimization, our knowledge base will be continuously improved to provide more accurate bidding matching services for enterprises.
### Configure API Key
In the `mcp - server.yaml` file, set the `jr - api - key` field to a valid API key.
### Integrate into MCP Client
On the user's MCP Client interface, add the relevant configuration to the MCP Server list.

View File

@@ -0,0 +1,67 @@
# Bid Tools MCP Server
## 君润标讯数字员工是什么
- 招投标是企业获得项目和客户的重要渠道,君润标讯数字员工能够提供全面、精准的标讯信息,帮助企业有效提升投标能力。
## 传统的标讯获取渠道存在什么问题
- 通过标讯平台获取标讯是过去常用的方式,但标讯来源多,一般标讯平台难以覆盖全,不同标讯平台的标讯存在大量重复,去重和筛选难
- 传统的标讯筛选仅通过关键词过滤,精准度不高;大量的标讯还需要依赖有经验的人员筛选,人力成本高
## 君润标讯数字员工能做什么
- 根据订阅的标讯信息每天通过邮件推送经过AI筛选后的精准标讯可以查看标讯的标题、详情、下载附件
- 基于AI的能力进行招标书的解析自动提取关键信息大幅提升投标决策的效率
- 基于历史标讯,提供未来的招标预测,提前布局
- 分析客户历史招标和中标情况,为制定投标策略提供支持
- 分析友商的历史投标和中标情况,获得目标客户清单
## 为什么选择君润标讯数字员工
### 标讯数据更全
- 与多家源头数据企业合作,涵盖政府标讯、第三方招标讯、企业自主网站标讯等,共计覆盖 10 万余个首发信息源站29 万余个采集频道地址1 万余个新媒体公众号源持续更新中每日更数万条招标、采购信息全网公开招标采购信息覆盖率可达98%以上(其余 2%由新资源和已有资源改版产生)
### 筛选结果更精准
- 基于AI大模型、数据训练及匹配技术相比传统的关键字搜索、规则判断精准度大幅提高
### 投标决策效率更高
- 历史中标情况分析、招标书一键解标,提取关键信息,有效辅助投标决策
## 数据源覆盖范围
| 站点类型 | 站点数量10万+ | 采集源地址数量29万+ | 信息发布量占比 |
|------------------|----------|------------|----------------|
| 企业招标 | - |- | 13.1183% |
| 电子采购平台 | - | - | 33.0614% |
| 金融银行 | - | - | 0.0757% |
| 高等教育 | - | - | 0.9645% |
| 教育局(网) | - | - | 0.0173% |
| 医院 | - | - | 0.8030% |
| 其他卫生机构 | - | - | 0.0167% |
| 人民政府 | - | - | 8.6457% |
| 公共资源中心 | - | - | 11.6248% |
| 政府采购中心 | - | - | 27.8139% |
| 工程交易中心 | - | - | 3.5826% |
| 行政服务中心 | - | - | 0.2761% |
![标讯治理流程](https://img.alicdn.com/imgextra/i1/O1CN01UMagYW28vlbrvgHAP_!!6000000007995-2-tps-557-262.png)
## 功能
- getBidlist根据关键字查询全网标讯信息。输入多个关键字返回招标查询结果。
- getBidinfo根据标讯ID查询标讯详情。输入标讯ID返回标讯详情。
- Email 企标匹配邮件通知服务(线下配置)根据企业的行业、业务范围、关键词群等等信息组合为企业专属画像再结合阿里百炼平台qwen3大模型深度思考每日标讯为企业匹配精准的标讯匹配服务。
## 使用教程
### 获取 apikey
1. 注册账号 [Create a ID](https://moonai-bid.junrunrenli.com?src=higress)
2. 发送邮件to: yuanpeng@junrunrenli.com 标题MCP 内容:申请标讯工具服务,并提供你的账号。
### 知识库(阿里百炼知识库)
1. 企业专属画像越明确,匹配效果越好。通过持续调优,我们的知识库将不断完善,为企业提供更精准的标讯匹配服务。
### 配置 API Key
`mcp-server.yaml` 文件中,将 `jr-api-key` 字段设置为有效的 API 密钥。
### 集成到 MCP Client
在用户的 MCP Client 界面,将相关配置添加到 MCP Server 列表中。

View File

@@ -0,0 +1,57 @@
server:
name: jr-ohris-bid-mcp
config:
apikey: ""
tools:
- name: getBidList
description: |+
根据关键字查询标讯列表。
- 输入关键字、开始时间、结束时间等信息。
- 支持可选的地区编码和当前页参数。
- 返回标讯列表。
args:
- name: keywords
description: 关键字
type: string
required: true
- name: cr_start_time
description: 开始时间,格式为 yyyy-MM-dd HH:mm:ss
type: string
required: true
- name: cr_end_time
description: 结束时间,格式为 yyyy-MM-dd HH:mm:ss
type: string
required: true
- name: areas
description: 地区编码
type: string
required: false
- name: page
description: 当前页
type: string
required: false
requestTemplate:
argsToUrlParam: false
url: https://agent-bid.junrunrenli.com/agent/tools/getBidList?jr-api-key={{.config.apikey}}
method: POST
headers:
- key: Content-Type
value: application/json
- name: getBidInfo
description: |+
根据关键字标讯ID查询详情。
- 输入标讯ID。
- 返回标讯详情。
args:
- name: id
description: 标讯ID
type: string
required: true
requestTemplate:
argsToUrlParam: false
url: https://agent-bid.junrunrenli.com/agent/tools/getBidInfo?jr-api-key={{.config.apikey}}
method: POST
headers:
- key: Content-Type
value: application/json