Houtini LM Token Saver

Example Use Cases

• 1

  Reduce token usage and costs by offloading bounded tasks from Claude Code to local or cloud LLMs.

• 2

  Automatic model discovery and routing to select the best model for specific tasks.

• 3

  Structured JSON output and performance tracking for delegated tasks.

Description

Houtini LM connects Claude Code to local LLM servers or OpenAI-compatible APIs to delegate well-defined tasks like code review, commit message generation, and format conversion. This reduces token usage by avoiding sending large source files to Claude, resulting in significant cost savings. It supports model discovery, automatic routing, structured JSON output, and performance tracking to optimize delegation and output quality.