Version 0.1.5-beta.6 (Prerelease)

v0.1.5-beta.6

fix: tokens omit empty

Version 0.1.5-beta.5 (Prerelease)

v0.1.5-beta.5

readme: update usage (#80)

Version 0.1.5-beta.4 (Prerelease)

Feat:

• claude more tool type support #63
• basic aiproxy http handler #64
• vertexai claude cache usage #68
• dashboard token usage #69
• group model config #70
• optimize log sql #71

Fix:

• add group token swag query docs #59
• rpm detail group can be empty #60
• dashboard rpm detail from cache #61
• vertex ai claude stream mode dosnot resp input usage #65

Version 0.1.5-beta.3 (Prerelease)

Feat:

• full swat params #58
• channel group sets support #57

Fix:

• price and usage typed #55
• check gemini has image #52

Version 0.1.5-beta.2 (Prerelease)

Feat:

• swagger docs #50
• docs: update listen readme #49

Version 0.1.5-beta.1 (Prerelease)

Feat:

• custom listen addr #48

Fix:

• xunfei returns the input modelname when no case matches for obtaining the model domain #46

Version 0.1.4

Feat:

• open router reasioning to reasoning_content #44
• rpm mem limit and limit header #43
• embed tiktoken #34
• cache token billing support #33
• grok api and error handler #32
• gemini image #31
• log query result onl #30
• baiduv2 rerank #26

Fix:

• exclude tests model #41
• update azure default api version #42
• dependabot no need push image #39
• parse pdf price #37
• clean modeltype and clean no used package #36
• need write status code when context done #28
• baidu and vertexai token expr #27
• claude handle balance error #23

Version 0.1.3

Fix:

• ali rerank model #21

Version 0.1.2

Fix:

• ollama tool call #19
• ali stt realtime #20

Version 0.1.1

Feat:

• aliyun image mirror #14
• vertexai claude think support #17
• channel designated from header Aiproxy-Channel: id #7 #11

Fix:

• miss record when prepare retry #13
• update ali model rpm #12
• vertexai key format #8