Gemini-OpenAI-Proxy

Bring OpenAI style API with low cost.

• Use Gemini models to complete OpenAI API chat and embedding calls.
• Use supertonic to implement tts-1 model.

Introduction

Gemini-OpenAI-Proxy is a proxy software. It is designed to convert OpenAI API protocol calls into Google Gemini Pro protocol, so that software using OpenAI protocol can use Gemini Pro model without perception.

If you're interested in using Google Gemini but don't want to modify your software, Gemini-OpenAI-Proxy is a great option. It allows you to easily integrate the powerful features of Google Gemini without having to do any complex development work.

Demo

Get api key from https://makersuite.google.com/app/apikey Deploy supertonic on huggingface or see github, set TTS_ENDPOINT in .dev.vars

curl -s http://localhost:8000/v1/chat/completions \
  -H "Authorization: Bearer $YOUR_GEMINI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
      "model": "gpt-4",
      "messages": [{"role": "user", "content": "Hello, Who are you?"}],
      "temperature": 0.7
      }'

curl http://localhost:8000/v1/audio/speech   -X POST   -H "Authorization: Bearer $YOUR_GEMINI_API_KEY"   -H "Content-Type: application/json"   -d '{ "input": "Hello secure world!!", "voice": "alloy" ,"model":"tts-1" }'   --output test.wav

curl -s http://localhost:8000/v1/chat/completions \
  -H "Authorization: Bearer $YOUR_GEMINI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "gpt-4-vision-preview",
  "messages": [
    {
      "role": "user",
      "content": [
        {
          "type": "text",
          "text": "What do you see in this picture?"
        },
        {
          "type": "image_url",
          "image_url": {
            "url": "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAADAAAAAnAgMAAAA0vyM3AAAACVBMVEX/4WwCAgF3aTMcpbzGAAAAa0lEQVR4nGOgAWB1QOYEIHFEcXKmhCBxQqYgcSLEEGymAFEEhzFAFYmTwNoA53A6IDmB1YETidPAiLBVFGgEgrNqJYIzNTQU4Z5QZA6QNQ3hGpAZcNegceBOADFQOQlQDhfQyUwLkPxKVwAABbkRCcDA66QAAAAASUVORK5CYII="
          }
        }
      ]
    }
  ],
  "stream": false
}'

Plan

• [x] /v1/chat/completions
  • [x] stream
  • [x] complete

Model Mappings

Run On Serverless

Cloudflare Workers

Copy main_cloudflare-workers.mjs to cloudflare-workers

Deno Deploy

Copy main_deno.mjs to deno deploy

Vercel

• Alternatively can be deployed with cli: vercel deploy
• Serve locally: vercel dev
• Vercel Functions limitations (with Edge runtime)

Run On Local

deno

install deno

curl -fsSL https://deno.land/install.sh | sh
source ~/.bashrc

build

deno task build:deno

run

deno task start:deno

node

node dist/main_node.mjs

bun

bun dist/main_bun.mjs

deploy to cloudflare workers/pages

deno task build:deno
mkdir -p cf
cp ./dist/main_cloudflare-workers.mjs ./cf/_worker.js
npx wrangler dev ./cf/_worker.js --name gemini-openai --compatibility-date 2025-10-04 --port 8000


npx wrangler deploy ./cf/_worker.js --name gemini-openai --compatibility-date 2025-10-04

npx wrangler pages deploy ./cf --project-name gemini-openai

docker

docker run -d -p 8000:8000 ghcr.io/zuisong/gemini-openai-proxy:deno
## or
docker run -d -p 8000:8000 ghcr.io/zuisong/gemini-openai-proxy:bun
## or
docker run -d -p 8000:8000 ghcr.io/zuisong/gemini-openai-proxy:node

Star History