Open source AI-powered job search.
Runs in your CLI. Your data, your machine.

Frequently asked

How does career-ops score job listings?

career-ops uses a rubric-guided LLM evaluation across five dimensions — match, north-star alignment, comp, cultural signals, red flags — producing a holistic 1.0–5.0 global score with citations to specific CV lines and JD requirements. Anything below 4.0 the agent recommends against applying. No closed-form formula, no spray-and-pray. The full rubric is published at career-ops.org/methodology.

Does career-ops apply to jobs for me?

It prepares every application right up to the click: it scans roles, scores each against your CV, and tailors a resume. Then it hands the decision back to you. You review and send each one yourself. That is deliberate: mass auto-apply burns your standing with recruiters and ATS systems, so career-ops removes the busywork and keeps the choice yours. More in Can an AI agent run your whole job search?

Is career-ops free? What is the business model?

career-ops is permanently free, MIT-licensed, and community-funded. There is no paid tier, no waitlist, no account, no telemetry, and no premium features. You clone the repo, configure your profile, and run the system locally with whichever AI coding CLI you already use. Sustainability comes from voluntary community patronage via GitHub Sponsors — not from premium tiers, paid features, or data. The maintainer has other paid work for income; sponsorship enables deeper focus on the project. See career-ops.org/sustain for details.

Where does my data live?

On your own machine, in plain files you own: your CV, profile, pipeline, and reports are local Markdown and YAML. career-ops runs entirely locally through your AI CLI: no account, no telemetry, nothing uploaded to a career-ops server. System updates never touch your data layer; that separation is the Data Contract. The only data that leaves your computer is whatever your chosen AI CLI sends to its own provider.

Who built career-ops?

career-ops was built by Santiago Fernández de Valderrama Aparicio — an Applied AI Operator with 16+ years building products, founder and operator of a Spanish phone-repair business (2009–2025) before exiting, and currently Head of Applied AI at Zinkee. He created career-ops in early 2026 to manage his own AI-era job search — 740 listings evaluated, one Head of AI role landed — and open-sourced it under MIT once he no longer needed it.

Is career-ops a Claude Code skill or a standalone tool?

career-ops is CLI-agnostic. It works with Claude Code, OpenCode, Codex, GitHub Copilot, and more — whichever AI coding agent the user already pays for. The skill files (modes/) live in the repo as plain markdown prompts; any agent that supports skill loading can invoke them. There is no Anthropic-specific dependency. Claude Code happens to be the most common runtime because of its skill loader, but the same modes run unchanged in the other CLIs.

How is career-ops different from resume checkers and auto-apply tools?

career-ops is open source and MIT-licensed, runs locally on your own machine through whichever AI coding CLI you already use, and publishes its full evaluation rubric. A human stays in the loop on every application; it never auto-submits. There is no account, no telemetry, and no subscription to career-ops itself; the only recurring cost is the AI CLI you choose. For honest, side-by-side comparisons with specific tools, see career-ops.org/compare.

What AI tools does career-ops work with?

Claude Code, Codex, OpenCode, Antigravity CLI, Grok Build CLI, Qwen, Kimi, and GitHub Copilot CLI — eight first-class CLIs (Gemini CLI is a legacy wrapper). The same mode files run on all of them. Each user picks the CLI that fits their existing subscription and cost preferences — career-ops never locks you to one provider. A typical job search runs on Claude Pro at $20/month, but the choice is yours.