career-ops vs ai-job-search
Two of the top projects in Assistants, side by side: who leads on each score part and who each one suits, then setup, license, activity and what each review found.
career-ops
Job-search agent that scores listings, tailors CVs and drafts applications
ai-job-search
Claude Code workflow that searches job boards and drafts tailored CVs and cover letters
Verdict
Both score 95; career-ops ranks higher on stars, 79,179 to 68,613. Of the 4 score parts we measure for both, career-ops leads on 1, ai-job-search on 1 and 2 are even.
- Adoptioncareer-ops leads by 2 points, 94 to 92: close to even.
- FreshnessEven at 100.
- Maintenanceai-job-search leads by 4 points, 97 to 93: close to even.
- Agent-readyEven at 85.
Pick career-ops if you are
Job seekers comfortable with a terminal and an AI coding CLI
- Drafts only: no auto-submit, and no mail transport in the code
- No telemetry or backend; CV goes only to the AI provider you choose
Pick ai-job-search if you are
Job seekers comfortable with Claude Code, git and LaTeX
- Drafter-reviewer pipeline compiles CV and cover letter PDFs with page-count and ATS checks
- Postings pass a deterministic integrity gate for hidden text and injected instructions
| What we compare | career-ops | ai-job-search |
|---|---|---|
| Score parts, out of 100 | ||
| Adoption | 94, widely used | 92, widely used |
| Freshness | 100, active | 100, active |
| Maintenance | 93, healthy | 97, healthy |
| Agent-ready | 85, ready | 85, ready |
| Facts from GitHub and the README | ||
| Stars | 79.2k | 68.6k |
| License | MIT (permissive) | MIT (permissive) |
| Last commit | Oct 2026 | Oct 2026 |
| Last release | Oct 2026 | Oct 2026 |
| Language | JavaScript | Python |
| Docker | Yes | No |
| GPU | Not needed | Not needed |
| arm64 or Apple Silicon | Not stated | Not stated |
career-ops
Runs inside an AI coding CLI (Claude Code, Codex, OpenCode, Qwen and others). You paste a job URL or description, and it checks the posting is live, scores fit against your CV, and generates an ATS PDF CV, cover letter and draft answers. It never submits applications or sends email, and keeps data on your machine.
Who it is for: Job seekers comfortable with a terminal and an AI coding CLI
Strengths
- Drafts only: no auto-submit, and no mail transport in the code
- No telemetry or backend; CV goes only to the AI provider you choose
- Covers evaluation, tailored PDF CV, cover letter, contact drafts, tracker and interview prep
- Portal scanner preconfigured for 100+ companies across Ashby, Greenhouse, Lever and Wellfound
Weaknesses
- Needs a separate AI CLI and provider; it does not run models itself
- README says early runs are rough until it learns your profile
- In-code check on CV rewording cannot catch every altered claim
- PDF output needs Playwright Chromium; non-Debian Linux needs manual libraries
- no GPU
- Docker + Compose
- Needs Node.js, An AI coding CLI (Claude Code, Codex, OpenCode, Qwen), Playwright Chromium (PDF generation)
- Models: Claude Code, Codex, OpenCode, Qwen, local models (see docs/RUNNING_ON_A_BUDGET.md)
ai-job-search
A Claude Code project you fork and fill in with your profile. Slash commands (/setup, /scrape, /rank, /apply, /interview, /outcome) search job portals, score postings against a fit framework, and draft a LaTeX CV and cover letter that a second reviewer agent critiques before final PDF compilation. Portal skills target Danish boards, plus LinkedIn and freehire.me.
Who it is for: Job seekers comfortable with Claude Code, git and LaTeX
Strengths
- Drafter-reviewer pipeline compiles CV and cover letter PDFs with page-count and ATS checks
- Postings pass a deterministic integrity gate for hidden text and injected instructions
- Tracks applications end to end: /outcome, /gmail-sync, /notion-sync, /html-report, /interview
- Portal CLIs are auto-discovered, and /add-portal scaffolds one for your own market
Weaknesses
- Requires Claude Code with a paid subscription or API credits; other agent tools only partly supported
- Most portal skills target Denmark; other markets rely on LinkedIn, freehire.me or a custom portal
- Needs Python 3.10+, Bun and a full LaTeX install, which is a heavy setup
- A public fork exposes personal data in tracked files; a private repo is needed
- no GPU
- Needs Claude Code, Python 3.10+, Bun, LaTeX (lualatex, xelatex), pypdf or poppler (optional)
- Models: Claude (via Claude Code)
More in Assistants
Building with these?
We help teams pick an open-source AI stack and run it in production. Client work never changes a rank.