Claude Code Multi-Model Proxy

Use Claude Code with any model from providers like GreenPT, Eurouter, OpenRouter, Together AI, Groq, and more. This lightweight proxy translates Anthropic Messages API to OpenAI Chat Completions API format, unlocking access to hundreds of models.

🎁 Featured Providers

Provider	Highlights	Get Started
GreenPT	Fast inference, great coding models (Qwen, DeepSeek, Llama)	Get $5 Free Credits →
Eurouter	EU-based infrastructure, DeepSeek models	Visit Eurouter →

┌─────────────────┐      Anthropic API      ┌─────────────────┐      OpenAI API      ┌─────────────────┐
│                 │ ────────────────────▶   │                 │ ──────────────────▶  │    Eurouter     │
│   Claude Code   │                         │     Proxy       │                      │    GreenPT      │
│                 │ ◀────────────────────   │                 │ ◀──────────────────  │   OpenRouter    │
└─────────────────┘                         └─────────────────┘                      └─────────────────┘

Features

✅ Full message translation — Anthropic ↔ OpenAI formats
✅ Streaming support — Real-time response streaming via SSE
✅ Tool/Function calling — Anthropic tools to OpenAI functions
✅ System prompts — Proper system message handling
✅ Multi-turn conversations — Full context preservation
✅ Model mapping — Map Claude tiers (Opus/Sonnet/Haiku) to any models
✅ Web configuration UI — Configure via browser at http://localhost:8080
✅ Docker support — Easy containerized deployment

Quick Start

Option 1: pip (Recommended)

pip install claude-code-multi-model-proxy

Then run with GreenPT (get free credits):

claude-code-proxy \
  --target-url https://api.greenpt.ai/v1 \
  --api-key YOUR_API_KEY \
  --default-model qwen3-coder-30b-a3b-instruct

Or with Eurouter:

claude-code-proxy \
  --target-url https://api.eurouter.ai/api/v1 \
  --api-key YOUR_API_KEY \
  --default-model deepseek/deepseek-chat-v3

Option 2: From Source

git clone https://github.com/robertkeus/claude-code-all-models.git
cd claude-code-all-models
pip install -r requirements.txt
python -m anthropic_openai_proxy.proxy --target-url https://api.eurouter.ai/api/v1 --api-key YOUR_API_KEY

Option 3: Docker

docker run -p 8080:8080 \
  -e TARGET_URL=https://api.eurouter.ai/api/v1 \
  -e API_KEY=your-api-key \
  -e DEFAULT_MODEL=deepseek/deepseek-chat-v3 \
  robertkeus/claude-code-all-models

Option 4: Docker Compose

# Set your API key
export GREENPT_API_KEY=your-api-key

# Run
docker-compose up -d

Usage with Claude Code

Step 1: Start the proxy

claude-code-proxy \
  --target-url https://api.eurouter.ai/api/v1 \
  --api-key YOUR_API_KEY \
  --default-model deepseek/deepseek-chat-v3

Step 2: Configure Claude Code

# Log out of any existing Anthropic account
claude logout

# Set environment variables
export ANTHROPIC_BASE_URL=http://localhost:8080
export ANTHROPIC_AUTH_TOKEN=proxy
export ANTHROPIC_API_KEY=""

# Run Claude Code
claude --model gpt-4

One-liner

claude logout && \
ANTHROPIC_BASE_URL=http://localhost:8080 \
ANTHROPIC_AUTH_TOKEN=proxy \
ANTHROPIC_API_KEY="" \
claude --model gpt-4

Configuration

Command-Line Options

Flag	Environment Variable	Default	Description
`--target-url`	`TARGET_URL`	`https://api.openai.com/v1`	OpenAI-compatible API base URL
`--api-key`	`API_KEY`	`""`	API key for target endpoint
`--default-model`	`DEFAULT_MODEL`	`gpt-4`	Default model name
`--opus-model`	`OPUS_MODEL`	(default)	Model for Claude Opus requests
`--sonnet-model`	`SONNET_MODEL`	(default)	Model for Claude Sonnet requests
`--haiku-model`	`HAIKU_MODEL`	(default)	Model for Claude Haiku requests
`--port`	`PORT`	`8080`	Proxy server port
`--host`	`HOST`	`0.0.0.0`	Host to bind to

Model Mapping

Claude Code uses different models for different tasks. You can map each tier to specific models on your provider:

claude-code-proxy \
  --target-url https://api.eurouter.ai/api/v1 \
  --api-key YOUR_KEY \
  --default-model deepseek/deepseek-chat-v3 \
  --haiku-model deepseek/deepseek-chat-v3 \
  --sonnet-model deepseek/deepseek-chat-v3 \
  --opus-model deepseek/deepseek-r1

Claude Tier	Use Case	Suggested Mapping
Haiku	Quick tasks, summaries, titles	Fast/cheap model
Sonnet	Main coding tasks	Good coding model
Opus	Complex reasoning	Best available model

Web Configuration UI

Access the configuration interface at http://localhost:8080 after starting the proxy:

Select your provider (GreenPT, OpenRouter, or custom)
Enter your API key
Choose models for each Claude tier
Save and copy the launch commands

Supported Providers

Provider	Target URL	Example Model
GreenPT ⭐	`https://api.greenpt.ai/v1`	`qwen3-coder-30b-a3b-instruct`
Eurouter ⭐	`https://api.eurouter.ai/api/v1`	`deepseek/deepseek-chat-v3`
OpenRouter	`https://openrouter.ai/api/v1`	`openai/gpt-4-turbo`
Together AI	`https://api.together.xyz/v1`	`meta-llama/Llama-3-70b-chat-hf`
Groq	`https://api.groq.com/openai/v1`	`llama-3.1-70b-versatile`
Fireworks	`https://api.fireworks.ai/inference/v1`	`accounts/fireworks/models/llama-v3-70b`
LM Studio	`http://localhost:1234/v1`	`local-model`
Ollama	`http://localhost:11434/v1`	`llama3`
OpenAI	`https://api.openai.com/v1`	`gpt-4-turbo`

💡 New to this? Get $5 free credits from GreenPT to try it out!

API Translation Reference

Endpoints

Anthropic (incoming)	OpenAI (outgoing)
`POST /v1/messages`	`POST /v1/chat/completions`
`POST /v1/messages/count_tokens`	(estimated locally)

Request Format Translation

Anthropic format (Claude Code sends this):

{
  "model": "claude-3-sonnet-20240229",
  "max_tokens": 1024,
  "system": "You are a helpful assistant.",
  "messages": [
    {"role": "user", "content": "Hello"}
  ]
}

OpenAI format (Proxy sends this):

{
  "model": "gpt-4",
  "max_tokens": 1024,
  "messages": [
    {"role": "system", "content": "You are a helpful assistant."},
    {"role": "user", "content": "Hello"}
  ]
}

Response Format Translation

OpenAI format (Provider returns this):

{
  "choices": [{"message": {"content": "Hi there!"}, "finish_reason": "stop"}],
  "usage": {"prompt_tokens": 10, "completion_tokens": 3}
}

Anthropic format (Proxy returns this):

{
  "id": "msg_abc123",
  "type": "message",
  "role": "assistant",
  "content": [{"type": "text", "text": "Hi there!"}],
  "stop_reason": "end_turn",
  "usage": {"input_tokens": 10, "output_tokens": 3}
}

Troubleshooting

"Auth conflict" error in Claude Code

You need to log out of Anthropic first:

claude logout

Connection refused

Make sure the proxy is running before starting Claude Code:

# Terminal 1: Start proxy
claude-code-proxy --target-url ... --api-key ...

# Terminal 2: Start Claude Code
ANTHROPIC_BASE_URL=http://localhost:8080 claude --model deepseek/deepseek-chat-v3

Model not found

Specify the exact model name that your provider supports:

claude --model qwen3-coder-30b-a3b-instruct

Streaming issues

If you experience streaming problems, the proxy automatically falls back to simulated streaming from non-streaming responses.

Development

Setup

git clone https://github.com/robertkeus/claude-code-all-models.git
cd claude-code-all-models

# Create virtual environment
python -m venv venv
source venv/bin/activate

# Install with dev dependencies
pip install -e ".[dev]"

Running Tests

pytest

Code Quality

# Format
black src/ proxy.py

# Lint
ruff check src/ proxy.py

# Type check
mypy src/

Publishing

To PyPI

# Install build tools
pip install build twine

# Build
python -m build

# Upload to PyPI
twine upload dist/*

To Docker Hub

# Build
docker build -t robertkeus/claude-code-all-models .

# Push
docker push robertkeus/claude-code-all-models

Architecture

anthropic-openai-proxy/
├── src/
│   └── anthropic_openai_proxy/
│       ├── __init__.py       # Package entry point
│       ├── proxy.py          # Main FastAPI application
│       ├── config.py         # Configuration & model mapping
│       ├── routes.py         # API route handlers
│       ├── converters.py     # Format translation logic
│       ├── streaming.py      # SSE stream handling
│       └── templates/
│           └── config.html   # Web configuration UI
├── public/
│   └── assets/               # Static assets
├── Dockerfile                # Docker build configuration
├── docker-compose.yml        # Docker Compose setup
├── pyproject.toml            # Python package configuration
└── requirements.txt          # Dependencies

Contributing

Contributions are welcome! Please see CONTRIBUTING.md for guidelines.

Fork the repository
Create a feature branch
Make your changes
Submit a pull request

Disclaimer

This project is not affiliated with, endorsed by, or associated with GreenPT, OpenRouter, Eurouter, Anthropic, or any other API provider mentioned in this documentation.

Use at your own risk. The authors and contributors are not responsible for any damages, data loss, API costs, or other issues that may arise from using this software. Always review your API provider's terms of service before use.

License

MIT License - See LICENSE for details.

Acknowledgments

Anthropic for Claude and Claude Code
FastAPI for the excellent web framework
All the OpenAI-compatible API providers

Name		Name	Last commit message	Last commit date
Latest commit History 3 Commits
public/assets		public/assets
src/anthropic_openai_proxy		src/anthropic_openai_proxy
.gitignore		.gitignore
CHANGELOG.md		CHANGELOG.md
CONTRIBUTING.md		CONTRIBUTING.md
Dockerfile		Dockerfile
LICENSE		LICENSE
README.md		README.md
claude-greenpt.sh		claude-greenpt.sh
docker-compose.yml		docker-compose.yml
pyproject.toml		pyproject.toml
requirements.txt		requirements.txt
start-greenpt.sh		start-greenpt.sh

Folders and files

Latest commit

History

Repository files navigation

Claude Code Multi-Model Proxy

🎁 Featured Providers

Features

Quick Start

Option 1: pip (Recommended)

Option 2: From Source

Option 3: Docker

Option 4: Docker Compose

Usage with Claude Code

Step 1: Start the proxy

Step 2: Configure Claude Code

One-liner

Configuration

Command-Line Options

Model Mapping

Web Configuration UI

Supported Providers

API Translation Reference

Endpoints

Request Format Translation

Response Format Translation

Troubleshooting

"Auth conflict" error in Claude Code

Connection refused

Model not found

Streaming issues

Development

Setup

Running Tests

Code Quality

Publishing

To PyPI

To Docker Hub

Architecture

Contributing

Disclaimer

License

Acknowledgments

About

Resources

Contributing

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages