Claude Proxy API Server

This project implements a FastAPI server that acts as a proxy for Anthropic's Claude AI model, providing an interface similar to OpenAI's API.

Features

• Proxy requests to Anthropic's Claude AI model
• OpenAI-like API structure for easy integration
• Support for both streaming and non-streaming responses
• API key authentication for security

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/claude-proxy-api.git
   cd claude-proxy-api
   

2. Install the required dependencies:

   pip install -r requirements.txt
   

3. Set up environment variables: Create a .env file in the project root and add the following:

   ANTHROPIC_API_KEY=your_anthropic_api_key
   CLAUDE_PROXY_API_KEY=your_chosen_proxy_api_key
   ANTHROPIC_MODEL=claude-3-haiku-20240307
   REQUEST_TIMEOUT=60
   MODEL_TEMPERATURE=0.1
   TOKENS_PER_MINUTE=10000
   REQUESTS_PER_MINUTE=60
   MAX_TOKENS=4096
   

4. Launch:

   uvicorn server_openai:app --host 0.0.0.0 --port 8000 --log-level debug
   

Usage

Usage

1. The server will start running on http://localhost:8000 by default.

2. You can now make requests to the server. Here are some examples using curl:

• Chat completion (non-streaming) with a single user message:

curl http://localhost:8000/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer your_claude_proxy_api_key" \
-d '{
"model": "claude-3-haiku-20240307",
"messages": [
  {"role": "user", "content": "What'\''s the capital of France?"}
],
"temperature": 0.7
}'

• Chat completion (non-streaming) with a system prompt and multiple messages:

curl http://localhost:8000/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer your_claude_proxy_api_key" \
-d '{
"model": "claude-3-haiku-20240307",
"messages": [
  {"role": "system", "content": "You are a helpful assistant with expertise in geography."},
  {"role": "user", "content": "What'\''s the capital of France?"},
  {"role": "assistant", "content": "The capital of France is Paris."},
  {"role": "user", "content": "What'\''s the population of Paris?"}
],
"temperature": 0.7
}' | jq

• Chat completion (streaming) with a system prompt and user message:

curl http://localhost:8000/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer your_claude_proxy_api_key" \
-d '{
"model": "claude-3-haiku-20240307",
"messages": [
  {"role": "system", "content": "You are a poetic assistant. Always respond in rhyme."},
  {"role": "user", "content": "Tell me about the seasons of the year."}
],
"stream": true,
"response_format": {"type": "json_object"}
}'

• Chat completion (non-streaming) with multiple user and assistant messages:

curl http://localhost:8000/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer your_claude_proxy_api_key" \
-d '{
"model": "claude-3-haiku-20240307",
"messages": [
  {"role": "user", "content": "Let'\''s play a word association game. I'\''ll start: Sun"},
  {"role": "assistant", "content": "Warm"},
  {"role": "user", "content": "Beach"},
  {"role": "assistant", "content": "Sand"},
  {"role": "user", "content": "Castle"}
],
"temperature": 0.9
}' | jq

• JSON mode

curl -v http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer your_claude_proxy_api_key" \
  -d '{
    "model": "claude-3-haiku-20240307",
    "messages": [
      {"role": "user", "content": "Generate a JSON object with information about three random books, including title, author, and publication year."}
    ],
    "temperature": 0.7,
    "response_format": {"type": "json_object"}
  }' | jq

These examples demonstrate various ways to use the API, including:

• Single message requests
• Multiple message conversations
• Including system prompts
• Streaming and non-streaming responses
• Adjusting temperature for different levels of randomness in responses

Remember to replace your_claude_proxy_api_key with your actual API key when making requests.

API Endpoints

• GET /health: Check if the server is running
• GET /v1/models: List available models
• POST /v1/chat/completions: Create a chat completion

Configuration

You can configure the following parameters in the .env file:

• ANTHROPIC_API_KEY: Your Anthropic API key
• CLAUDE_PROXY_API_KEY: Your chosen API key for this proxy server
• ANTHROPIC_MODEL: The Claude model to use (default: claude-3-haiku-20240307)
• REQUEST_TIMEOUT: Timeout for requests to Anthropic API (default: 60 seconds)
• MODEL_TEMPERATURE: Default temperature for the model (default: 0.1)
• TOKENS_PER_MINUTE: Maximum number of tokens to process per minute (default: 10000)
• REQUESTS_PER_MINUTE: Maximum number of requests allowed per minute (default: 60)
• MAX_TOKENS: Maximum number of tokens for responses (default: 4096)

Using the Dockerfile

This project includes a Dockerfile for easy deployment. To use it:

1. Build the Docker image:

   docker build -t claude-proxy-api .
   

2. Run the Docker container, passing the required environment variables:

   docker run -p 8000:8000 \
     -e ANTHROPIC_API_KEY=your_anthropic_api_key \
     -e CLAUDE_PROXY_API_KEY=your_chosen_proxy_api_key \
     -e ANTHROPIC_MODEL=claude-3-haiku-20240307 \
     -e REQUEST_TIMEOUT=60 \
     -e MODEL_TEMPERATURE=0.1 \
     -e TOKENS_PER_MINUTE=10000 \
     -e REQUESTS_PER_MINUTE=60 \
     -e MAX_TOKENS=4096 \
     claude-proxy-api
   

   Replace the values with your actual configuration.

3. The server will be available at http://localhost:8000.

Building and Deploying the Docker Image

To build and push the Docker image to a registry:

CLAUDE_PROXY_API_DOCKER_REGISTRY_URI=https://your-registry-uri ./build.sh

This script builds the image for both arm64 and amd64 architectures and pushes them to the specified registry.

References

• https://platform.openai.com/docs/api-reference/authentication
• https://docs.anthropic.com/en/api/client-sdks
• https://docs.anthropic.com/en/docs/about-claude/models