Auto-JobHunter
Auto-JobHunter is an industrial-grade, fully automated AI job application system designed for platforms such as BOSS Zhipin, 51job, and Liepin. The system executes a complete workflow starting with stealth scraping using Playwright to evade detection, followed by storing data in SQLite. It employs a LangGraph-based multi-agent system to evaluate job descriptions and dynamically rewrite resumes through a four-stage process of splitting, architecting, critiquing, and formatting. A proprietary veto filter allows users to define hard rules to immediately reject unsuitable roles like outsourced or sales positions, ensuring AI resources are spent only on high-quality matches. The platform supports a dual-channel data approach, synchronizing local SQLite databases with Feishu Bitable for cloud-based monitoring via a Next.js control panel featuring real-time logging. Key technical differentiators include DOM penetration to bypass frontend popups, physical profile isolation, and simulated human mouse trajectories. Whi