Houtini Lm
MCP server that saves Claude Code tokens by delegating bounded tasks to local or cloud LLMs. Works with LM Studio, Ollama, vLLM, DeepSeek, Groq, Cerebras.
- Transport
- Not stated
- Package
- —
- Registry id
- —
No install snippet on purpose. A working MCP config is a command, its arguments and an environment block — the last two are where API keys live, so this catalogue never stores them and cannot publish them. Follow the link above for the authors' own instructions.
Houtini LM is an MCP server that gives Claude Code extra tools to hand work to other models. Houtini LM supports local models in LM Studio, Ollama, vLLM, SGLang or llama.cpp, and cloud providers including OpenAI, OpenRouter, DeepSeek, Groq and Cerebras. Its code_task_files tool reads files itself so they never enter Claude's conversation.
It requires Node 22.5 or newer. It provides eight tools including chat, code_task, code_task_files, discover, and stats. It captures a thinking model's hidden reasoning.
It is available on npm and in the MCP registry at version 3.3.3. Every call can carry a model parameter to pin to one model, or leave routing to houtini-lm.
Summary drafted from the project's own website. Every sentence is backed by text on that page and was reviewed before publishing.