Headroom

Headroom (by Headroom Labs, founded by ex-Netflix engineer Tejas Chopra) is an open-source, local-first context compression layer that sits between coding agents and LLMs, reducing token usage before tool outputs, logs, files, and RAG chunks reach the model. The core open-source library is Apache 2.0 licensed and free; a paid desktop app and a nascent enterprise managed offering exist at very low price points targeting individual developers rather than large enterprises. Headroom Labs as an enterprise entity is explicitly described as still in stealth, meaning there is no documented enterprise security posture, SOC 2, HIPAA, or similar compliance, and no named enterprise customers. The token-savings headline (60-95%) is a vendor-stated ceiling on highly repetitive JSON; independent hands-on testing has seen figures closer to 26%, and the vendor's own documentation cautions that savings depend on workload shape. At v0.25.0 with the author acknowledging it is not battle-tested, this is a developer-grade tool to watch rather than a production enterprise buy today.