How to track job applications with a tracker you own

There are three ways to track a job search: a spreadsheet, a hosted app, or an open-source tracker that keeps the data in a file you own. They differ on one axis that matters more than the feature list: who owns the data. A spreadsheet is yours but manual. A hosted app is automated but lives on someone else's server. An open-source, file-based tracker gives you both, automation and ownership. career-ops is the third kind.

Ways to track a job search, compared

Most people start in a spreadsheet, move to a dedicated app when the spreadsheet gets tedious, and only later notice they have handed their entire search history to a vendor. The three approaches trade off along the same axes:

A spreadsheet wins on ownership and loses on effort: every status change is a manual edit, so it goes stale by week two. A hosted app wins on automation and loses on ownership: the moment you close the account, the history goes with it, and in the meantime your search lives in someone else's database. The open-source, file-based approach is the one that does not force the trade: the tooling does the updating, and the record is a plain file that stays on your machine.

The career-ops approach: your tracker is a file you own

career-ops keeps your whole pipeline in one plain-markdown file, data/applications.md, one row per application. It is a table you can open in any editor, diff in git, and grep from the terminal. There is no database and no server: the file is the source of truth.

The columns are fixed so the tooling and the dashboard agree on the shape:

| # | Date | Company | Role | Score | Status | PDF | Report | Notes |
| 12 | 2026-03-14 | Acme AI | Staff Engineer, Platform | 4.6/5 | Interview | ✅ | [12](../reports/012-acme-ai-2026-03-14.md) | onsite next week |

The # is the report number and the row's identity (it is never renumbered), Report links to the full evaluation, and Status holds exactly one label from the pipeline: Evaluated, Applied, Responded, Interview, Offer, Hired, plus the terminal states Rejected, Discarded, and SKIP. Because everything is a file, the history is versionable: git log data/applications.md is a timeline of your whole search.

Two things read and write that one file, and it helps to know which does what:

• Tracker mode is the writer. You talk to it in your AI CLI ("mark Acme as Applied", "what is overdue for follow-up?") and it edits the rows and shows you where everything stands.
• The Go dashboard is the reader. It is a keyboard-driven terminal view over the same data/applications.md. It is not a second database, it is a view.

One markdown file as the source of truth; tracker mode writes it, the dashboard visualizes it. Nothing about your search leaves your machine, and career-ops keeps zero telemetry either way.

Track your applications in practice

The tracker is not a separate tool you set up; it is a mode of the same system that scans and evaluates roles. Once you have run an evaluation, every listing is already a row. To see where everything stands, open your AI CLI in the career-ops folder and run:

/career-ops tracker

It reads data/applications.md and gives you the chat-side summary: what is active, what is waiting on a reply, what is overdue. For the full keyboard-driven board, launch the Go dashboard, which groups the same rows by stage. Either way you are looking at the file you own, and every status change is a line in your git history.

Full reference for the mode is in the tracker mode docs.

Frequently asked

What is the best way to track job applications? The best way is whatever you will actually keep updated, which usually means automation plus ownership. A spreadsheet is fully yours but you maintain it by hand. A hosted tracker automates the work but stores your history on someone else's server. An open-source, file-based tracker gives you both: the tool updates the records for you, and the records stay in a plain file on your machine you can read, edit, and version in git.

Is there an open source job application tracker? Yes. career-ops is an MIT-licensed, open-source job-search system whose tracker keeps every application in a plain-markdown file you own (data/applications.md), one row per role with its status, score, and report. Its tracker mode updates the file conversationally, and a Go dashboard reads the same file as a keyboard-driven view. Nothing is stored on a server, and there is no account or lock-in.

Can I keep my job search data private? Yes. With a file-based tracker your data never leaves your machine. career-ops keeps your applications, CV, and reports as local files with zero telemetry, so nothing is uploaded unless you push it somewhere yourself. That is the opposite of a hosted tracker, where your search history lives in the vendor's database.

Tracking is one stage of the pipeline. See how the whole thing runs human-in-the-loop, or start with the Quick Start.