- add support for
gemini-1.5-pro-vision-latest
andgemini-1.5-flash-latest
- Resolve CORS error: Update server configurations to enable Cross-Origin access.
- Add support for function calls in non-stream mode.
- Migrate to itty-router to reduce package size.
- Utilize the official Deno Docker image for improved consistency.
-
gemini-openai-proxy can now act as a reverse proxy for google gemini, which can be useful for people in region that don't have access to the google gemini
it will request
https://generativelanguage.googleapis.com
curl \ "http://localhost:8000/v1/models/gemini-pro:generateContent?key=$YOUR_GEMINI_API_KEY" \ --header 'Content-Type: application/json' \ --data '{"contents":[{"parts":[{"text":"Hello"}]}]}'
-
Service settings can now be passed via apikey, currently the first supported setting is
useBeta
, which can be set like soit will use
v1beta
version gemini api, https://ai.google.dev/docs/api_versionscurl http://localhost:8000/v1/chat/completions \ -H "Authorization: Bearer $YOUR_GEMINI_API_KEY#useBeta" \ -H "Content-Type: application/json" \ -d '{ "model": "gpt-3.5-turbo", "messages": [{"role": "user", "content": "Hello"}], "temperature": 0.7,"stream":true }'
- change default safe setting config #8
- fix can't pull docker image with no manifest #13
-
fix bun runtime sream model problem, add ployfill for
TextDecoderStream
bun does not support
TextDecoderStream
now oven-sh/bun#5648 -
Imporve document
- Add cloudflare workers support
- Reduce docker image size
- Imporve document
- Refined the handling of consecutive user messages
- Add system message support
- Add stream api support
- support gemini-pro api
- support gemini-pro-vision api