An AI-powered desktop app that scrapes 150+ job sources, scores fit against your resume, tailors per role, drafts outreach, and fills the application — you review and submit. Free. Open source. Runs on Mac, Windows, and Linux.
The app runs locally on your computer. Your data never leaves. It stops being a glorified job-board bookmark and actually moves the search forward.
Scrapes 150+ job sources daily — company career pages, ATS portals (Greenhouse, Lever, Ashby, Workday), and major boards. Dedupes across sources. Surfaces what's actually new since you last looked.
Scores every job against your master resume. Flags strong matches, hides obvious misses, explains the gap when the gap matters.
Finds the right person at the company — recruiter, hiring manager, mutual — and drafts the message + a cover letter for the role. You approve before anything sends.
Rewrites your resume per role from a single master document. Keeps bullets truthful — surfaces what's already there, doesn't fabricate.
Applied, screening, onsite, offer, ghosted. Every interaction in one place, threaded by company. No more "where was I with X?"
A chat sidebar that knows your resume, your pipeline, the JD on screen. Ask it anything; it answers in context.
Drop in your resume and a few target companies. The agents do the rest — you stay in the loop for every send.
Upload your master resume. Add target roles, comp band, locations.
Board agent populates your queue with scored, ranked jobs every morning.
One click → tailored resume, cover letter, referral draft, application filled. You read, edit, submit — nothing sends without you.
Every reply, screening, onsite lands in the tracker. No spreadsheet, no Notion, no drift.
The app runs locally. Your resume, applications, drafts, and contacts live on your machine — not in a vendor database waiting to be sold, breached, or shut down.
The repo goes public when the beta ships. Read it, audit it, fork it. MIT-licensed and no telemetry.
If yours isn't here, reply to the launch email and we'll add it.
Always free. The desktop app is MIT-licensed and runs on your computer. You bring your own API key — Anthropic, OpenAI, OpenRouter, or a local Ollama model — so you see and control the inference cost.
Cross-platform from day one — native builds for macOS (Apple Silicon + Intel), Windows, and Linux. Single download per platform, around 80MB. No Docker, no Node setup, no cloud account. Double-click to install.
The full pipeline, end to end:
Steps 3 and 8 are the human checkpoints. Every send stays under your control — nothing leaves the app without you signing off.
Your resume, applications, drafts, cover letters, contacts — all on your computer, in a local database. Nothing is uploaded, sold, or used to train models. The app talks to the AI provider you choose, via your API key. It talks to nobody else.
Yes. During onboarding the app picks the right job-board sources for where you live, so you're not downloading every scraper. Missing a local board? It's MIT-licensed — a contribution adds it for everyone. Comp normalisation + location matching cover the EU, UK, Canada, Australia, India, and more out of the box.
For now, no — and that's a deliberate choice, not a missing feature. Auto-apply trains recruiters to ignore you. The app surfaces, scores, tailors, drafts, and fills the form, but every submission needs your eye on it. Quality beats volume.
Bring your own key — Anthropic, OpenAI, OpenRouter, or a local Ollama model. You control the inference, you see the cost, you swap models when you want.
The same pipeline, polished. Experimental flags removed, edge cases handled, rough UI smoothed. Plus the Copilot — a chat sidebar that knows your resume, your pipeline, and the JD on screen. Beta members get the upgrade for free.
The repo is open from day one. The beta carries a handful of experimental features and feedback prompts so we learn what's working before defaults are locked. The official release retires the experiments, fixes the rough edges, and ships polish — same code base, always free.
Waitlist members get the private beta first and shape what makes the cut for the official release.