Loading...
Discovering amazing AI tools

Headroom compresses tool outputs, logs, files, and RAG chunks before they reach the LLM, cutting 60-95% of tokens while preserving answers.
Headroom compresses tool outputs, logs, files, and RAG chunks before they reach the LLM, cutting 60-95% of tokens while preserving answers.
Headroom is an open-source context-compression toolkit that shrinks everything an AI agent reads - tool outputs, logs, RAG chunks, files, and conversation history - before it reaches the LLM, achieving 60-95% fewer tokens with the same answers. It bundles several compression techniques: SmartCrusher for statistical JSON and array compression (70-90% on tool outputs), AST-aware code compression via tree-sitter, and text and log compression for search results, build logs, and diffs. Its Compress-Cache-Retrieve (CCR) approach is reversible: originals are never deleted, so the LLM can retrieve full content on demand. Headroom ships as a Python package and a TypeScript package (headroom-ai), an OpenAI- and Anthropic-compatible HTTP proxy, and an MCP server, so it can drop into existing stacks with little or no code change.
Compare Headroom: vs Agent-Reach · vs SkillSpector · vs LMCache · vs Fonda