PimEyes
Face search engine that finds other public photos of the same person across the web.

- Categories
- Image and Video AnalysisPeople Search
- Type
- Web app
- Price
- Freemium
- Setup
- Easy to use
PimEyes is a face search engine. You upload a photo (up to 15 MB). It looks for other pictures of that face on the open web and comes back with source links. The company says a search takes under five seconds and that it indexes billions of images.
The important limit is on the FAQ. It searches the open web and excludes social media and video platforms. If the only photo of someone is their Instagram, PimEyes is the wrong tool. If the face appeared on a news site, a blog, or a forgotten homepage, this is one of the engines that might still have it.
How it works in practice
The system is matching a face, not the whole frame. A different haircut, a different background, or a group photo can still hit. That is the point, and also the failure mode. Lookalikes exist. PimEyes publishes a high accuracy number on the marketing site. I treat a hit as a lead and I open the source page.
Paid plans remove CAPTCHA, un-blur results, and add monitoring. Open Plus is the entry tier (the site currently advertises up to 75 searches a day). PROtect adds alerts and result management. Advanced is the high-volume tier with more alerts. They say uploaded photos are not stored without permission and are not used to train models. GDPR and encryption claims live on the same page.
When I use it
A photo is the only identifier. A journalist’s byline picture that might have been reused. A person who does not have a useful name string. I do not use it as a dating background check, and I do not treat a match as identification.
FaceCheck.ID covers more of the social-and-news mix and is louder about crime-and-dating use cases. PimEyes is the more buttoned-up vendor. Neither is a Gem on this site. Face search is useful and I will pay for it on a case that starts with a picture. It is also the category most likely to waste your time on a cousin who looks similar.



