Happy slice of mcp pizzaMCP.pizza

headroom

MCP.Pizza Chef: headroomlabs-ai

Lets you cut the amount of information your AI assistant has to read — tool outputs, log files, documents, and search results — before it reaches the assistant, without changing the answers you get. It works alongside coding assistants such as Cursor, VS Code, Cline, and Goose, watching what they read and shrinking it automatically. Setting it up means installing a small program on your computer and connecting an account key; it is best suited to people comfortable installing developer tools rather than complete beginners.

Coding
Data

Use This MCP server To

Cut my AI coding assistant's bill by shrinking what it reads Work with huge log files without overwhelming my assistant Keep my AI's answers the same while using far less text See a live dashboard of how much I'm saving Pull back the original text any time I need it Feed large documents to my assistant without hitting size limits

README

  ██╗  ██╗███████╗ █████╗ ██████╗ ██████╗  ██████╗  ██████╗ ███╗   ███╗
  ██║  ██║██╔════╝██╔══██╗██╔══██╗██╔══██╗██╔═══██╗██╔═══██╗████╗ ████║
  ███████║█████╗  ███████║██║  ██║██████╔╝██║   ██║██║   ██║██╔████╔██║
  ██╔══██║██╔══╝  ██╔══██║██║  ██║██╔══██╗██║   ██║██║   ██║██║╚██╔╝██║
  ██║  ██║███████╗██║  ██║██████╔╝██║  ██║╚██████╔╝╚██████╔╝██║ ╚═╝ ██║
  ╚═╝  ╚═╝╚══════╝╚═╝  ╚═╝╚═════╝ ╚═╝  ╚═╝ ╚═════╝  ╚═════╝ ╚═╝     ╚═╝
              The context compression layer for AI agents

60–95% fewer tokens (for JSON data), 15-20% fewer tokens (for coding agents) · library · proxy · MCP · content-aware compressors · local-first · reversible

CI codecov PyPI npm Model: Kompress-v2-base License: Apache 2.0 Docs

Docs · Install · Proof · Agents · Discord · llms.txt

AI agents / LLMs: read /llms.txt here, or fetch the live index / full docs blob.

chopratejas%2Fheadroom | Trendshift

Headroom compresses everything your AI agent reads — tool outputs, logs, RAG chunks, files, and conversation history — before it reaches the LLM. Same answers, fraction of the tokens.

Headroom in action
Live: 10,144 → 1,260 tokens — same FATAL found.

What it does

headroom FAQ

Can I use this to make my coding assistant faster and cheaper to run?
Yes — headroom automatically shrinks the tool outputs, files, and search results your assistant reads before they reach it, so each conversation uses far less text while giving you the same answers.
Can I use this to work with large log files or data files without overwhelming my assistant?
Yes — it is built to compress large text and structured data, such as JSON, by a large amount before your assistant ever sees it, so oversized files stop being a problem.
Which apps and AI assistants does this work with?
It is built to work with coding tools such as Cursor, VS Code, Cline, and Goose, and it can also run alongside other AI assistants like Claude, OpenAI's ChatGPT, and Google's Gemini through its own connector.
Do I need an account or an API key to use it?
Yes, you will need an account key to connect it to your AI assistant before it can start compressing what the assistant reads.
How hard is this to set up?
This tool is aimed at people who are comfortable installing developer software — setup involves installing a small program on your computer and running a command-line install, so it is not a beginner-friendly one-click tool.
Will using this change the answers my AI assistant gives me?
No — it is designed to shrink what your assistant reads without changing the substance of the answers it gives back to you.
Can I get back the original, uncompressed text if I need it later?
Yes — it keeps a local copy of the original text so your assistant can retrieve the full version again if it is ever needed.
Does my data get sent to an outside server for this to work?
No — the compression happens locally on your own computer, so your files and conversations stay with you rather than being uploaded elsewhere.