Zero-dependency Node.js proxy with auto API key rotation for OpenAI/Gemini

3 months ago 2

A robust Node.js proxy server that automatically rotates API keys for Gemini and OpenAI APIs when rate limits (429 errors) are encountered. Built with zero dependencies and comprehensive logging.

  • 🔄 Automatic Key Rotation: Seamlessly switches to the next API key on 429 errors
  • 🎯 Multi-API Support: Works with both Gemini and OpenAI APIs simultaneously
  • 🔧 Flexible Base URLs: Use custom endpoints or default API servers
  • 📝 Detailed Logging: Track every request, rotation, and error with masked API keys
  • 🚀 Zero Dependencies: Pure Node.js with no external packages
  • 📁 File Upload Support: Handles multipart/form-data and binary uploads
  • 🛡️ Error Handling: Proper error responses and graceful failures
git clone <your-repo-url> cd openai-gemini-api-key-rotator

Copy the example environment file and add your API keys:

Edit .env:

# Required PORT=3000 # At least one of these is required GEMINI_API_KEYS=AIzaSyABC123...,AIzaSyDEF456...,AIzaSyGHI789... OPENAI_API_KEYS=sk-proj-abc123...,sk-proj-def456...,sk-proj-ghi789... # Optional - Custom base URL for all API calls # BASE_URL=https://your-custom-server.com

You'll see output like:

[CONFIG] Port: 3000 [CONFIG] Using default API endpoints [CONFIG] Found 3 Gemini API keys [CONFIG] Found 2 OpenAI API keys [GEMINI-ROTATOR] Initialized with 3 API keys [OPENAI-ROTATOR] Initialized with 2 API keys Multi-API proxy server running on port 3000 Available Gemini API keys: 3 Gemini endpoints: /gemini/v1/* and /gemini/v1beta/* Available OpenAI API keys: 2 OpenAI endpoints: /openai/v1/*
API Endpoint Pattern Example
Gemini /gemini/v1/* /gemini/v1/models/gemini-pro:generateContent
Gemini Beta /gemini/v1beta/* /gemini/v1beta/models/gemini-pro:generateContent
OpenAI /openai/v1/* /openai/v1/chat/completions

1. Generate Content (Text)

curl -X POST "http://localhost:3000/gemini/v1/models/gemini-2.5-pro:generateContent" \ -H "Content-Type: application/json" \ -d '{ "contents": [{ "parts": [{ "text": "Write a short poem about API key rotation" }] }] }'

2. Generate Content with System Instructions

curl -X POST "http://localhost:3000/gemini/v1/models/gemini-2.5-pro:generateContent" \ -H "Content-Type: application/json" \ -d '{ "system_instruction": { "parts": [{ "text": "You are a helpful coding assistant." }] }, "contents": [{ "parts": [{ "text": "Explain what API rate limiting is" }] }], "generationConfig": { "temperature": 0.7, "maxOutputTokens": 100 } }'
curl -X GET "http://localhost:3000/gemini/v1/models"
curl -X GET "http://localhost:3000/gemini/v1/models/gemini-2.5-pro"

5. Generate Content with Image (Base64)

curl -X POST "http://localhost:3000/gemini/v1/models/gemini-2.5-pro:generateContent" \ -H "Content-Type: application/json" \ -d '{ "contents": [{ "parts": [ { "text": "What do you see in this image?" }, { "inline_data": { "mime_type": "image/jpeg", "data": "base64_encoded_image_data_here" } } ] }] }'

6. Generate Content with Image (File Upload)

# First, encode your image to base64 IMAGE_DATA=$(base64 -i path/to/your/image.jpg) curl -X POST "http://localhost:3000/gemini/v1/models/gemini-2.5-pro:generateContent" \ -H "Content-Type: application/json" \ -d '{ "contents": [{ "parts": [ { "text": "Describe this image in detail" }, { "inline_data": { "mime_type": "image/jpeg", "data": "'$IMAGE_DATA'" } } ] }] }'

7. Generate Content with Multiple Images

curl -X POST "http://localhost:3000/gemini/v1/models/gemini-2.5-pro:generateContent" \ -H "Content-Type: application/json" \ -d '{ "contents": [{ "parts": [ { "text": "Compare these two images and tell me the differences" }, { "inline_data": { "mime_type": "image/jpeg", "data": "base64_image1_data_here" } }, { "inline_data": { "mime_type": "image/png", "data": "base64_image2_data_here" } } ] }] }'
curl -X POST "http://localhost:3000/openai/v1/chat/completions" \ -H "Content-Type: application/json" \ -d '{ "model": "gpt-4o-mini", "messages": [ { "role": "system", "content": "You are a helpful assistant." }, { "role": "user", "content": "Explain API key rotation in simple terms" } ], "max_tokens": 150, "temperature": 0.7 }'

2. Chat Completions with Streaming

curl -X POST "http://localhost:3000/openai/v1/chat/completions" \ -H "Content-Type: application/json" \ -d '{ "model": "gpt-4o-mini", "messages": [ { "role": "user", "content": "Write a haiku about programming" } ], "stream": true }'
curl -X GET "http://localhost:3000/openai/v1/models"

4. Text Completions (Legacy)

curl -X POST "http://localhost:3000/openai/v1/completions" \ -H "Content-Type: application/json" \ -d '{ "model": "gpt-4o-mini", "prompt": "The benefits of API key rotation are:", "max_tokens": 100, "temperature": 0.5 }'
curl -X POST "http://localhost:3000/openai/v1/embeddings" \ -H "Content-Type: application/json" \ -d '{ "model": "text-embedding-ada-002", "input": "API key rotation helps maintain service availability" }'

6. Vision API - Analyze Image with GPT-4o-mini

curl -X POST "http://localhost:3000/openai/v1/chat/completions" \ -H "Content-Type: application/json" \ -d '{ "model": "gpt-4o-mini", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "What do you see in this image?" }, { "type": "image_url", "image_url": { "url": "https://example.com/image.jpg" } } ] } ], "max_tokens": 300 }'

7. Vision API - Analyze Local Image (Base64)

# First, encode your image to base64 IMAGE_DATA=$(base64 -i path/to/your/image.jpg) curl -X POST "http://localhost:3000/openai/v1/chat/completions" \ -H "Content-Type: application/json" \ -d '{ "model": "gpt-4o-mini", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in detail" }, { "type": "image_url", "image_url": { "url": "data:image/jpeg;base64,'$IMAGE_DATA'" } } ] } ], "max_tokens": 500 }'

8. File Upload - Create Assistant File

curl -X POST "http://localhost:3000/openai/v1/files" \ -H "Content-Type: multipart/form-data" \ -F "file=@path/to/your/document.pdf" \ -F "purpose=assistants"

9. Audio Transcription - Upload Audio File

curl -X POST "http://localhost:3000/openai/v1/audio/transcriptions" \ -H "Content-Type: multipart/form-data" \ -F "file=@path/to/your/audio.mp3" \ -F "model=whisper-1"

10. Audio Translation - Upload Audio File

curl -X POST "http://localhost:3000/openai/v1/audio/translations" \ -H "Content-Type: multipart/form-data" \ -F "file=@path/to/your/audio.mp3" \ -F "model=whisper-1"

11. Text-to-Speech - Generate Audio

curl -X POST "http://localhost:3000/openai/v1/audio/speech" \ -H "Content-Type: application/json" \ -d '{ "model": "tts-1", "input": "Hello, this is a test of the text-to-speech API", "voice": "alloy" }' \ --output speech.mp3
Variable Required Description Example
PORT ✅ Yes Server port 3000
GEMINI_API_KEYS 🔶 Optional* Comma-separated Gemini API keys AIza...,AIza...
OPENAI_API_KEYS 🔶 Optional* Comma-separated OpenAI API keys sk-proj-...,sk-proj-...
BASE_URL ❌ No Custom base URL for all APIs https://api.example.com

*At least one API key type is required

Popular API Provider Examples

# OpenRouter (supports 100+ models including Claude, GPT-4, Llama, etc.) OPENAI_BASE_URL=https://openrouter.ai/api OPENAI_API_KEYS=sk-or-v1-your-key-here # Groq (ultra-fast inference for Llama, Mixtral, Gemma models) OPENAI_BASE_URL=https://api.groq.com/openai OPENAI_API_KEYS=gsk_your-groq-key-here # Together AI (open source models) OPENAI_BASE_URL=https://api.together.xyz OPENAI_API_KEYS=your-together-key-here # Anthropic Claude (direct) OPENAI_BASE_URL=https://api.anthropic.com OPENAI_API_KEYS=sk-ant-your-key-here # Use custom proxy or local server OPENAI_BASE_URL=https://your-proxy-server.com # OPENAI_BASE_URL=http://localhost:8080 # Use default endpoints (OpenAI official) # OPENAI_BASE_URL=

The server provides detailed logging for monitoring and debugging:

[CONFIG] Loading configuration from /path/to/.env [CONFIG] Port: 3000 [CONFIG] Using default API endpoints [CONFIG] Found 2 Gemini API keys [CONFIG] Found 3 OpenAI API keys [CONFIG] Gemini Key 1: [AIza...1234] [CONFIG] Gemini Key 2: [AIza...5678] [CONFIG] OpenAI Key 1: [sk-p...9012] [CONFIG] OpenAI Key 2: [sk-p...3456] [CONFIG] OpenAI Key 3: [sk-p...7890] [GEMINI-ROTATOR] Initialized with 2 API keys [OPENAI-ROTATOR] Initialized with 3 API keys [INIT] Gemini client initialized [INIT] OpenAI client initialized [REQ-abc123def] POST /openai/v1/chat/completions from 127.0.0.1 [REQ-abc123def] Proxying to OPENAI: /v1/chat/completions [OPENAI::sk-p...9012] Currently active API key (1/3) [OPENAI::sk-p...9012] Attempting POST /v1/chat/completions (attempt 1) [OPENAI::sk-p...9012] Rate limited (429) - rotating to next key [OPENAI::sk-p...9012] Key marked as failed (1/3 failed) [OPENAI-ROTATOR] Rotated from index 0 to 1 -> [OPENAI::sk-p...3456] [OPENAI::sk-p...3456] Currently active API key (2/3) [OPENAI::sk-p...3456] Attempting POST /v1/chat/completions (attempt 2) [OPENAI::sk-p...3456] Success (200) [REQ-abc123def] Response: 200
  1. Request Routing: Incoming requests are routed based on URL patterns
  2. Key Selection: The first available API key is selected
  3. API Call: Request is forwarded to the appropriate API endpoint
  4. Error Handling:
    • 429 (Rate Limited): Automatically rotates to next key and retries
    • Other Errors: Returns the original error response
  5. Exhaustion: When all keys are rate-limited, returns 429 with descriptive error
openai-gemini-api-key-rotator/ ├── src/ │ ├── config.js # Environment configuration │ ├── keyRotator.js # API key rotation logic │ ├── geminiClient.js # Gemini API client │ ├── openaiClient.js # OpenAI API client │ └── server.js # HTTP proxy server ├── index.js # Main entry point ├── package.json # Node.js dependencies ├── .env.example # Environment template └── README.md # This file

Made with ❤️ for developers who hate rate limits

Read Entire Article
  1. Homepage
  2. Technology
  3. Zero-dependency Node.js proxy with auto API key rotation for OpenAI/Gemini

Related

What predicts involuntary singlehood: effects of self-esteem/desirable traits

What predicts involuntary singlehood: effects of self-esteem...

18 minutes ago 0
Show HN: Wiki&Database for Escape from Duckov Game

Show HN: Wiki&Database for Escape from Duckov Game

19 minutes ago 0
If you've been burned by WireGuard meshes in real infra

If you've been burned by WireGuard meshes in real infra

23 minutes ago 0
Hostinger Web Hosting