freerouter 0.1.0

Free LLM Router Service based on LiteLLM

FreeRouter

🛠️ LiteLLM Configuration Management Tool - Automated Multi-Provider Configuration Generation

中文文档 | English

What is this?

FreeRouter is a configuration management tool for LiteLLM.

Core Features:

• 📋 Automatically fetch model lists from Provider APIs
• ⚙️ Generate LiteLLM config.yaml configuration files
• 🚀 One-command startup of LiteLLM service

Important:

• FreeRouter does not provide AI services; all APIs and routing are provided by LiteLLM
• Recommend reading LiteLLM Documentation first
• If you're familiar with writing LiteLLM configs manually, you may not need FreeRouter

Supported Providers

Provider	Type	Free
OpenRouter	Text, Vision, Multimodal	✅ Partially Free
iFlow	Text	✅ Fully Free
ModelScope	Text	✅ Free Quota
OAI	Any (OpenAI-compatible)	Depends
Ollama	Text, Vision	✅ Local Free
Static	Any	Depends

Free Providers:

• OpenRouter (https://openrouter.ai/) - 30+ free models (GPT-3.5, Gemini, Llama, etc.)
• iFlow (https://iflow.cn/) - Chinese free models (Qwen, GLM, DeepSeek, etc.)

Quick Start

1. Installation

pip install freerouter

Or from source:

git clone https://github.com/mmdsnb/freerouter.git
cd freerouter
pip install -e .

2. Initialize Configuration

freerouter init

3. Configure Providers

Edit .env to add API keys:

OPENROUTER_API_KEY=sk-or-v1-xxxxx
IFLOW_API_KEY=sk-xxxxx

Edit config/providers.yaml to enable services:

providers:
  # OpenRouter - Free models
  - type: openrouter
    enabled: true
    api_key: ${OPENROUTER_API_KEY}

  # iFlow - Chinese free models
  - type: iflow
    enabled: true
    api_key: ${IFLOW_API_KEY}

  # Ollama - Local models
  - type: ollama
    enabled: true
    api_base: http://localhost:11434

  # ModelScope - Chinese models (2000 calls/day free)
  - type: modelscope
    enabled: false
    api_key: ${MODELSCOPE_API_KEY}

  # OAI - Generic OpenAI-compatible API (auto-fetch models)
  - type: oai
    name: myservice
    enabled: false
    api_base: https://api.example.com/v1
    api_key: ${MYSERVICE_API_KEY}

  # Static - Custom service (single model, manual config)
  - type: static
    enabled: false
    model_name: gpt-3.5-turbo
    provider: openai
    api_base: https://your-api.com/v1
    api_key: ${YOUR_KEY}

4. Start Service

# Fetch models and start service
freerouter

# Or step by step
freerouter fetch   # Fetch model list
freerouter start   # Start service

Service will start at http://localhost:4000.

5. Use API

All API usage follows LiteLLM Documentation.

# List available models
curl http://localhost:4000/v1/models

# Call model (OpenAI-compatible API)
curl http://localhost:4000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer sk-1234" \
  -d '{
    "model": "google/gemini-pro",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

CLI Commands

freerouter              # Start service (default, auto fetch + start)
freerouter init         # Initialize config directory
freerouter fetch        # Fetch model list and generate config
freerouter start        # Start LiteLLM service
freerouter list         # List configured models
freerouter --version    # Show version
freerouter --help       # Show help

Config file search order:

1. ./config/providers.yaml (current directory)
2. ~/.config/freerouter/providers.yaml (user config)

License

MIT License - see LICENSE

Links

• GitHub
• LiteLLM
• OpenRouter
• iFlow
• Ollama

---

For issues, please visit Issues