Is It Training on Me?

Is It Training on Me? answers one question for 222 apps and services: does this company train AI models on what you put into it, and can you refuse? Every entry gets a grade from A (it cannot train on you) to F (trains by default, no way out) and rests on one sentence quoted verbatim from the company's own published policy, with a link to a dated archived copy of the page it came from. The site's build refuses to publish any quote that isn't word for word in the archive. Where a policy never answers the question, the entry says so and shows the search that was run, rather than guessing. What you get: - The grade, the exact clause, and the archived page behind it - Opt-out steps in the company's own words, where a switch exists, and what refusing costs you when it doesn't - Weekly re-reads of 341 policy documents, with a change feed showing the before and after when a clause moves - Right of reply: every company with a bad grade is written to, and every reply and every correction is published in full - My Stack: check the apps you use, with nothing sent to a server - A public JSON API, RSS, a weekly email, and a free badge any A- or B-graded company can display What makes it different: no opinions, no summaries, no tracking. Just the sentence, the date, and the copy. Made and run by one person in the UK, with no cookies, no analytics and no accounts.