// the find
DaKheera47/job-ops
job-ops: DevOps principles applied to job hunting. A self-hosted pipeline to track, analyze, and assist your application process
JobOps is a self-hosted job hunt dashboard that aggregates listings from 13+ job boards, runs AI scoring against your profile, rewrites your CV per-application, and auto-tracks replies via Gmail OAuth. It's for developers who are actively job hunting and want to cut the spreadsheet work without resorting to auto-apply bots.
The extractor architecture is genuinely well-designed — each job board is a separate TypeScript module with its own package.json, so adding a new source doesn't touch core code. Gmail tracking that auto-transitions application status from recruiter reply patterns is the kind of thing people usually build manually in Zapier. Multi-provider AI support (Ollama, OpenRouter, Codex, OpenAI, Gemini) means you're not locked into paying for an API key. The decision to not auto-apply is the right call — it's stated directly in the README with the actual reason, which is refreshingly honest.
The scraping approach for LinkedIn, Indeed, and Glassdoor is fundamentally fragile — these platforms actively block scrapers and change their HTML constantly, so expect breakage after any major site update. The AGPLv3 + Commons Clause combination is legally ambiguous and will make some companies' legal teams nervous about using it internally. CV rewriting quality is entirely dependent on your prompt engineering and chosen model — the README doesn't show what the actual output looks like, which is the one thing that matters most for this feature. At 3,300 stars it's clearly caught traction, but the project is still single-maintainer with commits from a small contributor pool, which is a real bus-factor risk for a tool people might depend on daily during a job search.