auto-identity-remove turns privacy cleanup into a repeatable personal-ops workflow

Privacy tooling usually breaks down in one of two ways: either it becomes a fear-based dashboard that never actually does the messy work, or it ships as a brittle script that assumes a patient technical user will babysit every edge case. auto-identity-remove caught my eye because it aims at a much more practical middle ground. It treats data broker removal as an operational workflow that should be repeatable, stateful, and boring in the best way.

The repo automates a genuinely unpleasant job: searching broker sites for your records, finding the right listing, submitting opt-out forms, solving CAPTCHAs when needed, and repeating the whole process on a schedule instead of pretending one pass is enough. The README positions it as a runner for hundreds of people-search and data broker sites, with persistent state so previously handled brokers are skipped until a re-check window expires. That sounds simple, but it is exactly the kind of product discipline many privacy tools miss. The hard part is not just form-filling. It is making a maintenance loop that people will keep trusting.

What makes the project more interesting than a generic automation script is its point of view on privacy. The docs are unusually explicit: personal data stays in a local config.json, broker submissions go directly to the brokers themselves, CAPTCHA solving only sends image bytes, and there is no telemetry or central service quietly collecting the very data the tool claims to protect. I like that because the repo understands the trust problem. A privacy tool that is vague about data flow is already starting from a credibility deficit.

There is also a strong product lesson in how the workflow is packaged. Setup is interactive, scheduling adapts across macOS, Linux, Windows, and Docker, state is persisted so the runner knows what was recently handled, and manual cases are surfaced cleanly instead of hidden behind false success. Even the notification model shows some taste: iMessage on macOS, webhooks elsewhere, plus dry-run and verify modes for users who want to inspect behavior before letting it loose. That is a much more mature shape than “clone this repo and good luck.”

I also appreciate that the project does not overclaim. The README is clear that submitted does not always mean deleted, some sites still require email confirmation, and verification is a separate concern. That honesty matters. Good automation products do not just maximize green checkmarks. They make failure states legible so users know what still needs attention.

For builders, the broader takeaway is that useful automation often looks less like AI magic and more like careful systems design around a recurring human headache. auto-identity-remove is compelling because it wraps persistence, scheduling, privacy boundaries, and escape hatches around a task most people would otherwise postpone forever. That is exactly the kind of small-but-real product thinking open source needs more of.

If you care about personal privacy tooling, or just want to study how to turn a messy manual process into a maintainable local-first utility, this repo is worth a look. It feels like a good example of solving the boring parts seriously enough that the whole experience becomes believable.

GitHub: https://github.com/stephenlthorn/auto-identity-remove Privacy model: https://github.com/stephenlthorn/auto-identity-remove/blob/main/docs/PRIVACY.md