Data Broker Opt-Out Tracker

A self-hosted, no-server tracker for getting your personal information off data-broker sites — the core value of paid services like Incogni (systematic opt-outs + re-monitoring), as a static web page you own.

Privacy story: nothing leaves your browser. No server, no accounts, no telemetry, no analytics. Your profile and progress live in your browser's localStorage and in JSON files you export yourself. The repo contains only public broker data.

Use it: open the GitHub Pages site, or clone and open index.html straight from disk (file:// works), or run any static file server.

The workflow (the app is just the scoreboard)

1. Set up a dedicated email alias for opt-outs (e.g. optouts.you@yourdomain), with a forwarding rule into your inbox. Verification links and compliance replies stay corralled, and you can kill the alias later.
2. Fill in your profile (name, city, state, email) — stored locally only.
3. Scan for your listings. Hit Start scan to step through every tier-1 site: each step gives you a direct search-my-name link (where the site's URL scheme allows), the opt-out page, and a site:-scoped web search, plus Found listing / No listing buttons that record the result and advance. Or hand the Copy scan prompt output to your AI agent for a read-only sweep (see below). Each broker row also has its own Find my listing link.
4. Do the cascading opt-outs first. The list is already in the right order:
   • PeopleConnect suppression center — one suppression covers Intelius, TruthFinder, Instant Checkmate, and US Search. ⚠️ Suppress, don't delete: their "right to delete" deletes your suppression account and the suppressions with it. Keep the account on file.
   • BeenVerified — clears sister sites (PeopleLooker, NeighborWho, Ownerly, NumberGuru).
   • SmartBackgroundChecks / PeopleFinders — same parent; one usually clears the other.
5. Work down the tier-1 list. Some (TruePeopleSearch especially) will have cleared on their own after the cascades — re-check before submitting.
6. Tier 2 brokers (Acxiom, LexisNexis, Epsilon, CoreLogic, Experian Marketing, Thomson Reuters CLEAR) have no public listings — use their privacy portals or send the generated deletion letter.
7. Expect 1–45 days per broker. Statutory response windows are typically 45 days; several brokers (Radaris, MyLife) are notorious for needing repeat rounds.
8. Recheck every 90 days. When you mark a broker Removed, the tracker schedules a recheck; overdue entries flip to Recheck due automatically (and show up in the next scan; a still-clear answer restarts the 90-day clock). Data brokers re-scrape public records — removal is a subscription you pay in attention.

Features

• Status tracking per broker (Not checked → Found listing → Request sent → Removed → Recheck due) with timestamps and notes
• Scan mode: guided step-through of every people-search site with per-site search links; results recorded as you go
• Dashboard counts and cascade-first ordering
• Deletion letter generator with a state dropdown citing the right statute (CCPA/CPRA, VCDPA, MODPA, TDPSA, …) and a mailto: deep link for email-method brokers
• JSON export/import of full state (also accepts the older prototype export format)
• Agent prompts for AI-assisted opt-outs (see below)
• brokers.json and statutes.json are standalone, PR-able data files

Using this with your AI agent

A web page cannot — and should not — command your browser agent directly: agents rightly treat page content as data, not instructions (that's what keeps prompt injection at bay). So this app generates prompts you hand to your own agent. Human-in-the-loop by design.

• Copy scan prompt: a read-only sweep — the agent searches every unchecked tier-1 site for your listing and reports found/not-found per site, no forms submitted. The perfect first agent task: low-stakes, no confirmations needed.
• Copy agent prompt (per broker): a self-contained, agent-agnostic prompt with the URL, the known gotchas, and safety conventions baked in. Paste into Claude, ChatGPT/Atlas, or any agent with browser access. "Open in Claude / ChatGPT" links prefill the same prompt.
• Copy session prompt: bundles everything currently in Found listing or Recheck due, in cascade order, into one batch prompt.
• Copy state as JSON: machine-readable status summary an agent can read (and tell you how to update).

What automates well: navigating opt-out flows, finding your listing, filling forms, drafting deletion emails, watching an inbox for verification links. What stays human: CAPTCHAs, the final click on any submission, and anything touching identity documents or payment. That second list is by design — these flows involve your legal identity, and the prompts explicitly instruct agents to stop and hand back control at those points.

Claude users: this repo ships a skill at skill/broker-optout/ that teaches Claude the full workflow (cascade order, per-broker gotchas, verification-email handling, status updates, quarterly rechecks) — copy the folder into ~/.claude/skills/ or your project's .claude/skills/. Other agents can be pointed at CLAUDE.md, which documents the same conventions in plain markdown.

Repo hygiene: no personal data, ever

• Profile + progress exist only in localStorage and files you export.
• .gitignore blocks *progress*.json and profile*.json from the first commit — git history is forever once pushed.
• sample-progress.json (fake data) is the only progress file that belongs in the repo; use it to demo the import flow.

Data files & contributing

• brokers.json — the broker list: opt-out URLs, method, gotcha notes, cascade relationships, lastVerified dates. PRs welcome; please update lastVerified and run node scripts/sync-data.mjs (regenerates the brokers.js/statutes.js mirrors that let the app run from file://).
• statutes.json — state → privacy-statute map for the letter generator.
• A weekly GitHub Action link-checks every opt-out URL and files an issue when links rot (the main failure mode of broker lists). Schema validation runs on every PR.

Attribution & license

Broker list and opt-out details adapted from the Big Ass Data Broker Opt-Out List by Yael Grauer and contributors, licensed CC BY-NC-SA 4.0. In keeping with that license, this project is non-commercial and released under CC BY-NC-SA 4.0 as well (LICENSE.md). Found a broken or changed opt-out flow? Please also contribute the correction upstream to BADBOOL.

Disclaimer

Not legal advice. Opt-out results vary by broker and reappearance is common; statutes and URLs change. This tool tracks your requests — it can't guarantee any broker honors them.