Why there's no account.
Job hunting while employed is a confidential activity. The standard model for a job tracker is a web app that asks for an email inbox and stores the search on a server no one but its operator controls. OnTrack doesn't do that.
What OnTrack does with your data
Applications, contacts, interview notes, and reminders are all written to a single SQLite file on your machine:
| macOS | ~/Library/Application Support/OnTrack/ontrack.db |
|---|---|
| Windows | %APPDATA%\OnTrack\ontrack.db |
Résumé and cover-letter files you import, plus archived posting PDFs, are copied into a folder next to that database — not left where you originally put them, so backing up one folder backs up everything.
The one network call
OnTrack makes exactly one kind of outbound request today, and only when you trigger it: fetching the page at a job posting URL you've saved, so it can be archived as text and a PDF. That request goes straight from your machine to the posting's own server — nothing is relayed through us, seen by us, or logged by us.
The experimental paste-to-parse feature can send a job description directly to Anthropic using an API key you supply. It is off by default, shows an estimated cost before each call, and only runs when you ask it to.
What we can't see
There is no server. Nothing here could see your data even if it wanted to — how many applications are tracked, whether the app was opened today, or that a given install exists at all.
API key handling
API keys are stored via the system keychain (macOS) or DPAPI (Windows), never enter the renderer process, and are never written to logs.
Open source
Not decided yet. If the source opens, the repository link goes here, front and centre — a public repo backs up everything on this page far better than a paragraph can.
If sync ever ships
End-to-end encrypted, with a key held only by you. Optional, off by default. If it can't be built that way, it won't be built.