A proxy server that enables Claude Code to work with OpenAI-compatible API providers. Convert Claude API requests to OpenAI API calls, allowing you to use various LLM providers through the Claude Code CLI.
- Full Claude API Compatibility: Complete
/v1/messagesendpoint support - Multiple Provider Support: OpenAI, Azure OpenAI, local models (Ollama), and any OpenAI-compatible API
- Smart Model Mapping: Configure BIG and SMALL models via environment variables
- Function Calling: Complete tool use support with proper conversion
- Streaming Responses: Real-time SSE streaming support
- Image Support: Base64 encoded image input
- Error Handling: Comprehensive error handling and logging
# Using UV (recommended)
uv sync
# Or using pip
pip install -r requirements.txtcp .env.example .env
# Edit .env and add your API configuration# Direct run
python start_proxy.py
# Or with UV
uv run claude-code-proxyANTHROPIC_BASE_URL=http://localhost:8082 claudeRequired:
OPENAI_API_KEY- Your API key for the target provider
Model Configuration:
BIG_MODEL- Model for Claude sonnet/opus requests (default:gpt-4o)SMALL_MODEL- Model for Claude haiku requests (default:gpt-4o-mini)
API Configuration:
OPENAI_BASE_URL- API base URL (default:https://api.openai.com/v1)
Server Settings:
HOST- Server host (default:0.0.0.0)PORT- Server port (default:8082)LOG_LEVEL- Logging level (default:WARNING)
Performance:
MAX_TOKENS_LIMIT- Token limit (default:4096)REQUEST_TIMEOUT- Request timeout in seconds (default:90)
The proxy maps Claude model requests to your configured models:
| Claude Request | Mapped To | Environment Variable |
|---|---|---|
| Models with "haiku" | SMALL_MODEL |
Default: gpt-4o-mini |
| Models with "sonnet" or "opus" | BIG_MODEL |
Default: gpt-4o |
OPENAI_API_KEY="sk-your-openai-key"
OPENAI_BASE_URL="https://api.openai.com/v1"
BIG_MODEL="gpt-4o"
SMALL_MODEL="gpt-4o-mini"OPENAI_API_KEY="your-azure-key"
OPENAI_BASE_URL="https://your-resource.openai.azure.com/openai/deployments/your-deployment"
BIG_MODEL="gpt-4"
SMALL_MODEL="gpt-35-turbo"OPENAI_API_KEY="dummy-key" # Required but can be dummy
OPENAI_BASE_URL="http://localhost:11434/v1"
BIG_MODEL="llama3.1:70b"
SMALL_MODEL="llama3.1:8b"Any OpenAI-compatible API can be used by setting the appropriate OPENAI_BASE_URL.
import httpx
response = httpx.post(
"http://localhost:8082/v1/messages",
json={
"model": "claude-3-5-sonnet-20241022", # Maps to BIG_MODEL
"max_tokens": 100,
"messages": [
{
7565
"role": "user", "content": "Hello!"}
]
}
)This proxy is designed to work seamlessly with Claude Code CLI:
# Start the proxy
python start_proxy.py
# Use Claude Code with the proxy
ANTHROPIC_BASE_URL=http://localhost:8082 claude
# Or set permanently
export ANTHROPIC_BASE_URL=http://localhost:8082
claudeTest the proxy functionality:
# Run comprehensive tests
python src/test_claude_to_openai.py# Install dependencies
uv sync
# Run server
uv run claude-code-proxy
# Format code
uv run black src/
uv run isort src/
# Type checking
uv run mypy src/claude-code-proxy/
├── src/
│ ├── main.py # Main server
│ ├── test_claude_to_openai.py # Tests
│ └── [other modules...]
├── start_proxy.py # Startup script
├── .env.example # Config template
└── README.md # This file
- Async/await for high concurrency
- Connection pooling for efficiency
- Streaming support for real-time responses
- Configurable timeouts and retries
- Smart error handling with detailed logging
MIT License