All tools
AI coding toolsUpdated 2 days ago

Headroom

By Headroom Labs

Apache-2.0 context compression layer, around 72,400 GitHub stars, that shrinks tool outputs, logs, file reads, and RAG chunks before they reach the model while leaving error lines byte for byte intact, with 21 to 57 percent token reductions in its published measurements; it runs on your own machine as a Python or TypeScript library, a local proxy, an MCP server, or a one-command wrap around an existing coding agent, and caches the originals so the model can pull full text back on demand.

Best for

  • cutting the token bill on agents that read large logs or JSON
  • fitting long tool output into a context window without losing the error
  • adding compression through a proxy with no application code changes
  • wrapping an existing agent such as Claude Code, Codex, or Cursor
  • keeping prompt content local instead of sending it to a compression service
Recent changes
  • Sep 16, 2026Added to the coding category: Apache-2.0 context compression layer, roughly 72,400 GitHub stars and about 18,700 weekly npm installs of headroom-ai, that compresses tool outputs, logs, file reads, and RAG chunks locally before they reach the model and caches originals for retrieval. No other catalogue entry covers token reduction at the context boundary.

Other AI coding tools