Quick start

1. Build your profile

careeros start is a guided onboarding that produces your profile.yaml — your facts, targets, and constraints. It opens by asking for your CV (optional; type skip to answer questions instead).

$ careeros start

Your profile is a source of truth. Every resume bullet CareerOS ever selects is a verbatim copy of something written here, so it's worth getting right. You generate it once and edit it rarely.

2. Discovery is already set up

Nothing to add here. The default source reads a free, open dataset (no API key, no secrets.env entry, no signup) covering 65 real ATS platforms, updated daily. It filters itself from the profile.yaml you just built.

Want the optional paid Fantastic Jobs source too (e.g. for its LinkedIn/YC/Wellfound coverage)? Real credentials live in .careeros/secrets.env, which is gitignored — config files hold only the name of an environment variable, never the key itself.

FANTASTIC_API_KEY=your-developer-fantastic-jobs-key

3. Run the daily pipeline

$ careeros daily

This runs the full pipeline: discover, normalize, dedupe, constraints, gate, evaluate, threshold, artifacts, application answers, and finally the report and Sheet. Deterministic stages cost nothing; the two reasoning stages spend tokens only on jobs that earned them.

4. Open your Sheet

Apply-tier jobs (score of 4.0 or higher) get a full row with a tailored resume, cover letter, and drafted application answers. Consider-tier jobs (3.5 to 4.0) get a row with the score and a one-line reason. You read the Sheet, top down, and apply to what's worth it.

Going deeper on one job

These run on demand, never as part of the daily loop:

$ careeros prep <job-id>    # a deep interview-prep report
$ careeros apply <job-id>   # draft answers for one job's form