career-ops-main

A local-first AI cockpit for your job search. Bring your own model — a free local Ollama or any API key — and run the whole hunt from one calm screen: discover roles, evaluate fit, tailor your résumé to a real LaTeX PDF, draft answers and cover letters, and track everything through to the interview.

Your résumé and keys never leave your machine.

Studio is a wrapper, not a fork. It drives the career-ops scripts you already have (scan.mjs, ollama-eval.mjs, openai-eval.mjs, the CV/LaTeX pipeline) and only ever writes the same user-layer files the CLI uses (cv.md, config/profile.yml, portals.yml). It never touches career-ops system files, so update-system.mjs keeps working. No framework, no build step — vanilla HTML/CSS/JS over a zero-dependency Node server.

Why it's different

• Bring your own model. Local Ollama (free, private, offline) or any OpenAI-compatible API — OpenAI, Anthropic, OpenRouter, Groq, DeepSeek, LM Studio, vLLM. Keys stay in studio/.local/ (gitignored).
• Local-first. No account, no cloud, no telemetry leaves the box. Your CV is a file on your disk.
• It does the work before you sit down. Scan on a schedule, evaluate in the background, auto-prepare high-fit matches — then just decide.
• It never fabricates and never auto-submits. Everything is grounded in your résumé; missing facts become [ADD: …]; the submit button is always yours.