legard/claude-code-proxy repository overview

⁠Claude Code Proxy

A proxy server that enables Claude Code CLI to work with OpenAI-compatible API providers. Convert Claude API requests to OpenAI API calls, allowing you to use various LLM providers (OpenAI, Azure OpenAI, local models) through the Claude Code interface.

⁠Features

• Full Claude API compatibility (/v1/messages endpoint)
• Multiple provider support (OpenAI, Azure OpenAI, Ollama, any OpenAI-compatible API)
• Smart model mapping via environment variables
• Complete tool use/function calling support
• Streaming responses with SSE
• Image input support (base64 encoded)
• Custom HTTP headers injection

⁠Quick Start

docker run -d \
  --name claude-proxy \
  -p 8082:8082 \
  -e OPENAI_API_KEY="your-openai-key" \
  -e BIG_MODEL="gpt-4o" \
  -e SMALL_MODEL="gpt-4o-mini" \
  legard/claude-code-proxy:latest

Or with docker-compose:

services:
  proxy:
    image: legard/claude-code-proxy:latest
    ports:
      - "8082:8082"
    env_file:
      - .env

⁠Configuration

Required environment variables:

• OPENAI_API_KEY - Your API key for the target provider

Optional configuration:

• BIG_MODEL - Model for Claude Opus requests (default: gpt-4o)
• SMALL_MODEL - Model for Claude Haiku requests (default: gpt-4o-mini)
• OPENAI_BASE_URL - Custom API endpoint for OpenAI-compatible providers
  • Default: https://api.openai.com/v1
  • Examples:
    • Azure OpenAI: https://YOUR_RESOURCE.openai.azure.com/openai/deployments/YOUR_DEPLOYMENT
    • Ollama: http://localhost:11434/v1
    • Other compatible APIs: Any OpenAI-compatible endpoint

See full configuration guide⁠

⁠Use with Claude Code

ANTHROPIC_BASE_URL=http://localhost:8082 claude

⁠Links

• GitHub Repository⁠
• Documentation⁠
• Issues⁠

⁠Supported Platforms

• linux/amd64
• linux/arm64