LLM Cache Proxy is a FastAPI-based application that serves as a caching layer for OpenAI's API. It intercepts requests to the OpenAI API, caches responses, and serves cached responses for identical requests, potentially reducing API costs and improving response times.

# Run the container
docker run -p 9999:9999 \
  -v $(pwd)/data:/app/data \
  ghcr.io/so2liu/llm-cache-server:latest

OR

# Run the container with custom API key
docker run -p 9999:9999 \
  -e OPENAI_API_KEY=your_api_key_here \
  -e OPENAI_BASE_URL=https://api.openai.com/v1 \
  -v $(pwd)/data:/app/data \
  ghcr.io/so2liu/llm-cache-server:latest

The proxy is now available at http://localhost:9999

Use /cache/chat/completions for cached requests

Use /chat/completions for uncached requests

• Caches responses from OpenAI's API
• Supports both streaming and non-streaming responses
• Compatible with OpenAI's chat completion endpoint
• Configurable via environment variables
• Dockerized for easy deployment
• Persistent cache storage

• Python 3.12+
• Docker (optional, for containerized deployment)

1. Clone the repository:

   git clone https://github.com/so2liu/llm-cache-server.git
   cd llm-cache-server
   

2. Install the required packages:

   pip install -r requirements.txt
   

Set the following environment variables:

• OPENAI_API_KEY: Your OpenAI API key
• OPENAI_BASE_URL: The base URL for OpenAI's API (default: https://api.openai.com/v1)

You can set these in a .env file in the project root.

1. Start the server:

   python -m app.main
   

2. The server will be available at http://localhost:9999

Recommended: Use pre-built image from GitHub Container Registry

docker run -p 9999:9999 \
  -e OPENAI_API_KEY=your_api_key_here \
  -e OPENAI_BASE_URL=https://api.openai.com/v1 \
  -v $(pwd)/data:/app/data \
  ghcr.io/so2liu/llm-cache-server:latest

Or build locally:

1. Build the Docker image:

   docker build -t llm-cache-server .
   

2. Run the container with persistent storage:

   docker run -p 9999:9999 \
     -e OPENAI_API_KEY=your_api_key_here \
     -e OPENAI_BASE_URL=https://api.openai.com/v1 \
     -v $(pwd)/data:/app/data \
     llm-cache-server
   

   This command mounts a data directory from your current working directory to the /app/data directory in the container, ensuring that the cache persists between container restarts.

• /chat/completions: Proxies requests to OpenAI's chat completion API without caching
• /cache/chat/completions: Proxies requests to OpenAI's chat completion API with caching

Both endpoints accept the same parameters as OpenAI's chat completion API.

To run the application in verbose mode, use the --verbose flag:

python -m app.main --verbose

Contributions are welcome! Please feel free to submit a Pull Request.

This project is licensed under the MIT License - see the LICENSE file for details.