A job search tool that keeps the whole search in one place

It runs on your machine

The whole application — index, accounts, resumes, pipeline — is one SQLite file in the project’s data directory. There is no analytics script, no tracking pixel and no third-party session. Back it up by copying a file; delete everything by removing it.

What is stored

• Your email and a scrypt hash of your password — never the password itself.
• Profile preferences, resume text and the skills parsed out of it.
• Saved, hidden and viewed jobs, plus your applications, notes and contacts.
• Outgoing email, written to a local outbox instead of being sent anywhere.

How AI is used

AI is optional and off by default — no provider ships configured, so on a fresh install briefs, cover letters, tailoring and interview prep all come from deterministic engines that run offline. Add a key — Anthropic, any OpenAI-compatible host, or a local Ollama — and only the text needed for that one request leaves the machine, cached locally so the same job is never sent twice. Even then the model may only return edits to bullets you wrote: it never emits a resume, and a guard in plain code rejects any rewrite that introduces a number, a name, a date or a credential your own bullet did not already contain.