Commit Graph

32 Commits

Author SHA1 Message Date
Kent Dong
ab419efda4 fix: Fix the incorrect reasoning content concat logic in ai-proxy (#1842) 2025-03-07 10:33:45 +08:00
Kent Dong
49617c7a98 feat: Unify the SSE processing logic (#1800) 2025-02-25 11:00:18 +08:00
Kent Dong
fabc22f218 feat: Support transforming reasoning_content returned by Qwen to OpenAI contract (#1791) 2025-02-21 17:32:02 +08:00
澄潭
49aad4152c Supports completions API & support config openai baseUrl through openaiCustomUrl (#1765) 2025-02-18 09:57:48 +08:00
澄潭
2dbde1833f ai proxy support passthrough path when api name is unknown (#1754) 2025-02-13 21:22:43 +08:00
pepesi
a84a382f1d feature: allow ai-proxy to forward standard AI capabilities that are … (#1704) 2025-02-12 15:23:44 +08:00
Se7en
ba74f4bbb9 fix: baidu api issue (#1685) 2025-01-16 21:42:43 +08:00
Se7en
579c986915 feat: retry failed request (#1590) 2024-12-26 18:30:50 +08:00
澄潭
39c007d045 optimize ai proxy (#1603) 2024-12-19 16:22:35 +08:00
rinfx
2f5709a93e qwen bailian compatible bug fix (#1597) 2024-12-17 16:57:31 +08:00
StarryNight
2a200cdd42 AI proxy return unified status in header phase (#1588) 2024-12-16 18:41:38 +08:00
pepesi
d026f0fca5 feat: ai-proxy support dashscope-finance (#1554) 2024-12-05 11:48:09 +08:00
Se7en
e705a0344f fix: qwen stream issue (#1564) 2024-12-03 13:10:47 +08:00
Se7en
d24123a55f feat: implement apiToken failover mechanism (#1256) 2024-11-16 19:03:09 +08:00
rinfx
f51408d7ff add bailian support (#1319)
Co-authored-by: Kent Dong <ch3cho@qq.com>
2024-09-18 16:35:13 +08:00
Yang Beining
5a87031c0e Fix error due to incompatible top_p range (#1250) 2024-08-26 18:28:43 +08:00
韩贤涛
7054f01a36 feat: Adapt to the Qwen multimodal model generation API (#1221) 2024-08-22 18:42:16 +08:00
rinfx
8fa1224cba support qwen compatible mode (#1205) 2024-08-15 18:52:49 +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
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
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
Kent Dong
5a2c6835f7 feat: Support embeddings API for Qwen in the ai-proxy plugin (#1079) 2024-07-08 19:37:08 +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
Kent Dong
ffee7dc5ea fix: Accommodate the incomplete function name in the initial event from Qwen (#1045) 2024-06-19 13:48:20 +08:00
Kent Dong
bdd050b926 fix: Fix tool_calls compatibility issues with LobeChat (#1009) 2024-05-30 19:20:48 +08:00
澄潭
ca1ad1dc73 Fixed the issue where an empty string system prompt would be set when enabling file id. (#999) 2024-05-24 16:27:02 +08:00
澄潭
e09edff827 Fix the issue with multiple system prompts when using qwen-long file id mode. (#994)
Signed-off-by: johnlanni <zty98751@alibaba-inc.com>
2024-05-24 14:43:34 +08:00
Kent Dong
76b5f2af79 feat: Enhance the feature of ai-proxy plugin (#976) 2024-05-22 20:30:46 +08:00
Kent Dong
559a109ae5 feat: Refactor Qwen stream event processing workflow (#939) 2024-05-15 11:43:45 +08:00
Kent Dong
333f9b48f3 feat: Add an AI-Proxy Wasm plugin (#921)
Co-authored-by: 澄潭 <zty98751@alibaba-inc.com>
2024-05-14 17:00:12 +08:00