Search code, repositories, users, issues, pull requests...

Our vision is to provide a comprehensive routing solution for the entire AI ecosystem, supporting language models, image generation, video generation, and embedding models. We provide interfaces compatible with OpenAI's Chat Completions API, as well as the new Responses API and Anthropic's Messages API, enabling effortless integration with any applications that use these APIs.

Note: LMRouter Core focuses specifically on request routing and model proxying. Advanced features like comprehensive API key management, user administration, billing systems, and analytics are handled by separate projects in the LMRouter ecosystem (currently in active development).

This demo server simply proxies all requests to OpenRouter while exposing both OpenAI Chat Completions and Anthropic Messages APIs. It's perfect for connecting Claude Code to any OpenRouter model without any additional setup on your side. It's also running under BYOK (Bring Your Own Key) mode, so you need to provide your own API keys for OpenRouter.

Recently, it's popular to use qwen/qwen3-coder, moonshotai/kimi-k2, openai/gpt-oss-120b, and openai/gpt-5[-mini] (gpt-5 requires BYOK on OpenRouter) as the model for Claude Code, but feel free to use any model you want.

Leave a Comment