Reads the posting, writes only what's true, keeps score of your search.
Both paths need an AI coding CLI. Either one works:
- Claude Code:
npm install -g @anthropic-ai/claude-code(claude.com/claude-code) - Codex:
npm install -g @openai/codex(github.com/openai/codex)
On an Apple Silicon Mac: download the signed, notarized app from the
latest release, or
brew install --cask codeswhat/tap/careerrat. You still need one AI CLI
installed.
On any other platform, or if you'd rather use the terminal: install with
npm (requires Node.js 24 or newer, check with node -v), then run it:
npm install -g careerrat
careerrat start claude # or: careerrat start codexEither path sets up your workspace, opens the local app at
http://localhost:7777, and hands you off to the agent. From there you just talk to it, and watch it work:
each step shows up as a small activity line, reading your resume, searching the
web, writing, instead of narrated text. No dashboard to learn first, no settings
to get right before it's useful.
- Let it onboard you. A conversation, not a form: it asks what roles you want, what you'll accept, what you won't, and the real work you've done, then builds your profile from the answers. Want to kick the tires first? Say "set me up with a quick sample profile."
- Paste a job posting, a link, or the sample in
examples/sample-jobs/, and say "evaluate this." You'll get a verdict: keep it or cut it, how well it actually fits, whether the money works, and what to do next. All from a real read of the posting, not a keyword match. You'll see it fetch and read the posting as activity lines before the verdict lands. - Say "write a résumé and cover letter for this." It builds them from your own evidence and refuses to invent anything you didn't tell it.
- Paste a recruiter email and say "draft a reply." It writes the reply and remembers the thread, so the next one has context.
- Open
http://localhost:7777and watch the job land, move through your funnel, and pick up history as you go. Every step in the chat shows as a plain-language activity line, an icon plus a label like "Searching the web" or "Reading files," with a spinner that settles once it's done. The assistant only speaks up to ask a question or hand you a result.
One first-run thing that looks broken but isn't: before you've onboarded,
careerrat doctor will report your setup is incomplete and list candidate/*.yml
files to create. That's expected. Onboarding fills them in.
Activity lines streaming while CareerRat researches market comp.
careerrat next # the one thing worth doing next
careerrat doctor # check your setup is healthy
careerrat update # pull the latest code; your data is untouchedIf the local app is already running, the update relaunches that recorded CareerRat process on the updated code. Unrelated processes are never stopped; CareerRat picks another loopback port instead.
The local app comes up with careerrat start. To run it on its own:
careerrat tracker # snapshot tracker.json for recovery
careerrat tracker-dev # serve http://localhost:7777 with live reloadUseful flags on start: --no-agent (workspace + local app only),
--no-dashboard, --agent <name> (override with a compatible agent command),
--port <n> (default 7777).
git clone https://github.com/CodesWhat/careerrat
cd careerrat
npm install
npm run hooks:install
npm link
careerrat start claudeCareerRat is a job-search workspace that runs on your own machine. Tell it what you're actually looking for, once. It reads real job postings and tells you which ones are worth your time, writes applications from things you've genuinely done, drafts your recruiter replies, preps you for interviews, and keeps track of where everything stands, so you're not rebuilding that picture from memory.
No account, no CareerRat server, no telemetry. CareerRat never phones home, and your files stay on your machine. What does go out: your AI CLI talks to its own provider to do the work, same as any other task you'd give it, and the app fetches public resources like job postings and company logos from the services that host them. The packaged desktop app also checks GitHub once a day for a newer release and shows an in-app notice, nothing more; it never downloads or installs anything on its own, and you can turn that off in Settings. See privacy for the details.
Most job tools match keywords, then spray a hundred applications and hope one sticks. CareerRat won't write a single line of a cover letter until it has read the whole posting and checked it against what you said you want: your comp floor, your location, your dealbreakers. Jobs that don't clear that bar, it tells you to skip, and says why.
And it won't lie for you. Every claim in a tailored résumé traces back to something you told it about your own work. If you didn't do it, it doesn't get written, full stop.
CareerRat is an agent runtime, not a form-filling script. The CLI sets up the workspace and serves the local app, but the job-search work happens inside your own agent, reading a set of skills that spell out how each step gets done. That's why you talk to it in plain language instead of memorizing subcommands, and why the first run just detects which AI CLI you have and gets out of the way.
Activity lines mid-run: "Using the research-comp skill," "Searching the web."
The rule underneath all of it: no tailoring, no applying, until the job has passed a real read of the posting. Titles and keywords are triage, not truth.
Same skills for anyone. A nurse, an engineer, and a driver each answer onboarding their own way and get the same loop back. (The mascot's a rat named Paul. He doesn't do the reading, but he's why the app looks the way it does.)
CareerRat covers the whole search, not just the writing. Twenty-eight skills, grouped by what you're actually trying to do:
- Search setup: turns your targeting into real sources, boards, and company career pages worth watching, and lets you tune or import them by hand.
- Sourced sweep: scans everything configured, dedupes, drops dead links, and coarse-triages every new posting so you see what's worth a closer look.
- Company discovery: finds companies likely to be hiring your kind of role from your own company thesis, checks each one is real, and proposes adding it.
- Board research: looks for new job boards worth watching in your field and proposes adding them.
- Full-posting evaluation: reads the whole posting and checks it against your comp floor, location, and dealbreakers before anything gets written. This is the gate every tailor or apply run has to clear first.
- Company health: checks layoff risk, hiring momentum, financials, sentiment, and leadership stability, and scores the company healthy, watch, or risky.
- Company research: pulls together a cited brief on a company across six angles, for vetting and for interview prep.
- Comp research: benchmarks the market rate for a role and location, so the comp check has something real to compare your offer against.
- Gap coaching: when a job lands at "worth a look, but" with named fit gaps, works out an honest plan to close them, or tells you straight there isn't one.
- Honest tailoring: résumés, cover letters, and short answers built only from things you've actually done, with a check that blocks anything half-finished.
- Screening answers: drafts one-off application-question answers grounded in your profile and evidence, and remembers your standard disclosures so you're never asked twice.
- Filling and submitting: fills the portal form for you, including LinkedIn's one-click Easy Apply, and defaults to letting you hit submit.
- LinkedIn tune-up: diffs your profile against what you're targeting and proposes honest rewrites, headline through Featured, with a preview before anything gets written back.
- Recruiter comms: drafts replies, follow-ups, scheduling, and negotiation, and keeps the whole thread on record.
- Scheduling: reads proposed times, checks your timezone and calendar, and drafts a clear availability reply.
- Calendar holds: writes interviews, assessments, and deadlines to Apple Calendar, Google Calendar, or Outlook.
- Reading your inbox and DMs: pulls recruiter replies and status changes out of Mail, Gmail, Outlook, and LinkedIn or Wellfound messages, so the tracker reflects what's actually happening.
- Status sync: reads your ATS dashboards and normalizes whatever label they use into one vocabulary.
- Outcome tracking: records what happened, learns the pattern for your role family, and tells you when your results say the strategy needs a rethink.
- Warm contacts: finds likely recruiters and hiring-team contacts at companies you're tracking, for you to review before reaching out.
- Interview prep: packets tailored to who you're talking to, built from your real work and the job itself.
- Live comp coaching: scripts and rehearses the actual negotiation with you while you're in the room, then debriefs after and folds the lesson back in.
Dashboard: stat cards, funnel, active pipeline from sourced through offer, table / board / calendar views, Tokyo Night and Gruvbox themes, light or dark.
- Roadmap: version themes and what's next
- Sources strategy: how job sources get curated
- Architecture and AGENTS.md: how the skills and the agent contract fit together
Bugs and feature requests: GitHub Issues.
Questions, ideas, and show-and-tell: GitHub Discussions.
Chat: CodesWhat Discord.
