中文简介

AI Job Search 是一个跑在你自己机器上的 AI 求职申请框架（基于 Claude Code）：评估招聘帖、按岗位定制简历、写求职信、准备面试，全流程一条龙。2026 年 3 月公开，已有 4.9 万 Star，MIT 开源。最硬核的是作者亲测：他本是地球物理学家，2025 年底被裁后用这套流程投了 69 份定制申请、拿到 20 个一面、签了一份 AI 工程师 offer（2026 年 6 月入职）。

• 核心命令：/setup 建档案（读你 documents/ 里的简历 PDF、领英导出、证书，或面谈补全）、/scrape 搜职位并按匹配度排序、/apply <url> 一条龙（评估匹配度→起草简历+求职信→第二个 agent 评审→修改定稿）；
• 扩展命令：/rank 批量打分、/interview 针对性面试准备（含模拟面试）、/outcome 记录投递结果、/upskill 技能差距热力图+学习计划、/html-report 离线数据看板，还能同步到 Notion / 读 Gmail 跟进状态；
• 语言国家无关：核心流程不绑定国家，职位搜索技能默认是丹麦市场（Jobindex 等），但设计上就是让你换成本地招聘网站的；
• 前置要求：Claude Code（需 Pro/Max 订阅或 API 额度）、Python 3.10+、Bun、LaTeX（简历用 lualatex 编译）。

重要提醒：公开 fork 会把你的个人资料（姓名、联系方式、工作经历、薪资期望）写进被追踪的文件——作者明确建议用私有仓库 + upstream 的方式，不要直接 fork。求职中的朋友值得一看。

English Summary

AI Job Search is an AI job application framework that runs on your machine (built on Claude Code): evaluate postings, tailor your CV per role, write cover letters, prep interviews — the full pipeline. Public since March 2026, 48.5K stars, MIT licensed. The strongest credential: the author, a geophysicist laid off in late 2025, ran his own search through it — 69 tailored applications, 20 first interviews, one signed AI-engineer contract (started June 2026).

• Core commands: /setup builds your profile (reads CV PDFs, LinkedIn exports, diplomas in documents/, or interviews you), /scrape searches portals ranked by fit, /apply <url> runs the pipeline (fit evaluation → draft CV + cover letter → reviewer agent critiques → revise);
• More commands: /rank batch-scores postings, /interview stage-specific prep with mock interviews, /outcome tracks results, /upskill skill-gap heatmap + learning plan, /html-report offline dashboard, plus Notion and Gmail sync;
• Language- and country-agnostic: the core workflow isn’t tied to any market; portal skills default to Denmark (Jobindex etc.) but are designed to be swapped for your local boards;
• Prerequisites: Claude Code (Pro/Max subscription or API credits), Python 3.10+, Bun, LaTeX (CV compiles with lualatex).

Important: a public fork writes your personal data (name, contact, work history, salary expectations) into tracked files — the author explicitly recommends a private repo with this as upstream instead of forking. Worth a look if you’re job hunting.

链接 / Links

• 项目主页 / Repository: https://github.com/MadsLorentzen/ai-job-search