<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:sy="http://purl.org/rss/1.0/modules/syndication/" xmlns:media="http://search.yahoo.com/mrss/"><channel><title>Headroom on 陈少文的网站</title><link>https://www.chenshaowen.com/tags/headroom/</link><description>Recent content in Headroom on 陈少文的网站</description><generator>Hugo -- gohugo.io</generator><language>zh</language><copyright>&amp;copy;2016 - {year}, All Rights Reserved.</copyright><lastBuildDate>Fri, 03 Jul 2026 00:00:00 +0000</lastBuildDate><sy:updatePeriod>weekly</sy:updatePeriod><atom:link href="https://www.chenshaowen.com/tags/headroom/atom.xml" rel="self" type="application/rss+xml"/><item><title>Headroom：让 AI 编程助手更省 Token</title><link>https://www.chenshaowen.com/blog/headroom-reduce-ai-agent-tokens.html</link><pubDate>Fri, 03 Jul 2026 00:00:00 +0000</pubDate><atom:modified>Fri, 03 Jul 2026 00:00:00 +0000</atom:modified><guid>https://www.chenshaowen.com/blog/headroom-reduce-ai-agent-tokens.html</guid><description>1. Headroom 是什么 Headroom 是一款面向 AI Agent 的上下文压缩层。它在 LLM 收到请求之前，对工具输出、日志、RAG 片段、文件内容和对话历史做内容感知压缩，把 token 用量压到原来的 5%～40%，同时尽量保持回答质量不变。 简单说，它解决的是另一头的问题：CodeGraph 帮 agent 更</description><dc:creator>微信公众号</dc:creator><category>AI</category><category>MCP</category><category>Cursor</category><category>Headroom</category><category>研发</category></item></channel></rss>