Getting Started

Prerequisites

• Python 3.13+ (see pyproject.toml for the exact constraint)
• uv package manager (recommended) or pip
• An API key for your chosen provider (OpenAI, Azure, Ollama, etc.)
• Claude Code CLI installed (npm install -g @anthropic-ai/claude-code)

Install Dependencies

# Using uv (recommended)
uv sync

# Or using pip
pip install -r requirements.txt

Configure Your Provider

Copy the example environment file and set your API key and model names:

cp .env.example .env

Edit .env with your OpenAI credentials:

OPENAI_API_KEY="sk-your-openai-key"
BIG_MODEL="gpt-4o"
MIDDLE_MODEL="gpt-4o"
SMALL_MODEL="gpt-4o-mini"

For Azure OpenAI, Ollama, or other providers, see the Configuration Reference.

Start the Proxy

python start_proxy.py

The proxy starts on http://localhost:8082 by default.

Connect Claude Code

In a separate terminal, point Claude Code at the proxy:

ANTHROPIC_BASE_URL=http://localhost:8082 claude

That’s it — Claude Code now sends requests through the proxy to your chosen provider.

Verify the Setup

Check that the proxy is running:

curl http://localhost:8082/health

Run the built-in test script to confirm end-to-end connectivity:

python src/test_claude_to_openai.py