You heard about our shutdown
Here is the comeback.

2019: the app that was pulled

Like Patrol launched as an iOS app in 2019. It watched the accounts a user followed and reported their activity: what they liked, who they followed, when. In October 2019 Instagram removed its own Following activity tab and sent the app a cease-and-desist; in November Apple removed it from the App Store. The coverage called it a stalking app, and the criticism was fair. Reporting a person's likes to someone else, minute by minute, was not something people had agreed to. We are not going to argue with that history. We are going to be precise about what is different now.

What changed

Today's LikePatrol reads one thing: the following list of an account, the same list anyone can open by tapping Following on a profile. It does not read likes. It does not report activity. It does not notify anyone about anything a person does. Instead of surveillance, the product is analysis. It samples the public posts of the accounts on that list and describes what they are about, using a fixed taxonomy of content signals, so that a parent, coach, brand or researcher can understand the environment an account has chosen without reading hundreds of profiles by hand. A digest shows which accounts were followed and unfollowed, again from public snapshots, again with the same analysis attached. The report describes the content around a person. It never describes the person.

What we will never do

• Ask for anyone's password or log into any account.
• Read direct messages, stories, drafts, or location.
• Show likes, comments, or view history.
• Notify or contact the account being analyzed.
• Run facial recognition or any biometric analysis.
• Sell data, use it to train models, or share it with advertisers.
• Act as a background check. Our reports are not consumer reports and must not be used for employment, housing, credit, or insurance decisions.

How the analysis works

For every account on a following list that is not private, the pipeline collects a sample of public posts and profile text, and a language model answers a narrow question: is this kind of content present, yes or no, against a human-curated taxonomy. Tags are grouped into exposure dimensions informed by public-health and media-literacy research. The public content is used in the moment and discarded; only the derived signals are kept to render the report. Full details are in the FAQ and the privacy policy.

Who we are

A small independent team of engineers and researchers. We are not affiliated with, endorsed by, or operated by Meta, Instagram, TikTok, or any social platform, and we have no special API access to any of them. The free preview is open in a list of countries that we keep in configuration, not in code, and the product is built to operate within the laws of the territories where it is offered.