Awesome MCP › Other MCP Servers

headroomlabs-ai/headroom

⭐ 71937 Python added to this list on 2026-06-29 repository created 2026-01-07

Headroom is an innovative context compression layer designed to optimize AI agent and LLM interactions by significantly reducing token consumption. It processes various inputs such as tool outputs, logs, RAG chunks, files, and conversation history before they reach the language model, achieving token savings of 60-95% without compromising answer quality. The project offers multiple integration methods: as a Python/TypeScript library for in-app use, as a proxy for zero-code-change integration, and as an agent wrapper for specific coding agents like Claude Code or Cursor. Crucially, Headroom functions as an MCP (Model Context Protocol) server, providing `headroom_compress`, `headroom_retrieve`, and `headroom_stats` functionalities for any MCP client. Its architecture includes a `ContentRouter` for content-aware compression using specialized algorithms like `SmartCrusher` for JSON, `CodeCompressor` for AST, and `Kompress-base` for text. The `CacheAligner` stabilizes prefixes to improve LLM KV cache hits, and the `CCR` (Content-Consistent Reversibility) component ensures that original data can be retrieved on demand, making the compression reversible. Additional features include cross-agent memory for shared context, `headroom learn` for mining failed sessions and generating corrections, and output token reduction to trim LLM responses. Headroom emphasizes local-first operation, ensuring data privacy and control.

https://github.com/headroomlabs-ai/headroom

agentaianthropicclaude-codecompressioncontext-engineeringcontext-windowcursorfastapilangchainllmmcpopenaiprompt-engineeringproxypythonragtoken-optimizationtokenstypescriptreversible-compressioncode-compressionjson-compressiontext-compressiontoken-reduction