Facial Recognition Search: How PimEyes and FaceCheck.id Work

What can face-search engines actually find? Upload a photo to a tool like PimEyes or FaceCheck.id, and within seconds you can get a list of other places online where that same face appears — old social posts, news photos, dating profiles, anything the tool's crawler has indexed. They don't hand you a name and home address on a silver platter, but they hand you enough breadcrumbs that a motivated person often can. That gap between "technically anonymized" and "practically identifiable" is the whole story here.

• How reverse face-search actually works
• PimEyes
• FaceCheck.id
• Yandex Images
• The stalking and doxxing risk
• The Harvard I-XRAY project
• Opting out and removing yourself
• Which Ban the Bots page do you actually need?
• Frequently asked questions

How reverse face-search actually works

These tools work in three steps: crawl, index, match. A crawler scrapes publicly accessible images from across the web — social media, news sites, forums, blogs, anywhere a photo with a visible face sits unprotected. Each detected face gets converted into a mathematical representation, a faceprint, and stored in a searchable index. When you upload a query photo, the tool converts that face into the same kind of representation and returns the closest matches from its index, ranked by similarity.

None of this requires the person in the photo to have agreed to anything. If your face is in a publicly visible photo anywhere the crawler reaches, it can end up in the index regardless of whether you ever signed up for the service doing the crawling.

PimEyes

PimEyes is the best-known commercial face-search engine and the one most frequently cited in privacy reporting. It markets itself partly as a tool for people to check where their own photos are being used without permission, but nothing technically stops anyone from uploading a photo of someone else. Free searches show limited results; a paid subscription unlocks the full set of matched source pages, which is where the real risk lives — a stranger with $30 and a screenshot of your face can see a fuller picture of your online footprint than you might expect.

FaceCheck.id

FaceCheck.id operates on a similar model to PimEyes: upload a face, get a ranked list of visually similar faces found across the indexed web, with links back to source pages. It has marketed itself partly toward use cases like verifying a dating match's identity or checking whether a photo has been used in a scam, but the underlying capability — and the underlying risk of misuse against someone who never consented — is the same as any other face-search engine.

Yandex Images

Yandex, the Russian search engine, isn't marketed as a dedicated face-search product the way PimEyes and FaceCheck.id are — it's a general reverse-image search tool. But it has a long-running reputation among privacy researchers and journalists for producing unusually accurate face matches, particularly for people with any social media presence in Russia or the wider CIS region, where Yandex's crawling and image indexing is especially deep. That makes it a go-to secondary tool for anyone doing a serious face search, even outside its home region.

The stalking and doxxing risk

The core danger with face-search tools isn't the technology in isolation — it's what it enables when combined with other easily available data. A single face match to an old social media post can lead to a username, which leads to a hometown, which leads to a workplace, which leads to enough to show up somewhere in person. Each individual piece of information might be "public," but face search collapses the time and effort needed to assemble them into one place, which is exactly what makes it useful for stalking and harassment as well as for its stated legitimate uses.

This risk falls disproportionately on people who are already more likely to be targeted for harassment — journalists, activists, abuse survivors trying to stay hidden from a former partner, and public-facing professionals whose photos are already widely circulated.

The Harvard I-XRAY project

The clearest real-world demonstration of this risk came in late 2024, when Harvard students AnhPhu Nguyen and Caine Ardayfio built a project they called I-XRAY. It combined Meta's Ray-Ban Meta smart glasses with PimEyes and public-record lookup tools, feeding a live camera stream into a pipeline that could identify strangers walking by and surface their name, address, and phone number, all without the person ever knowing their face had been scanned.

The project was widely reported as a proof of concept rather than a released product — the creators said their goal was to demonstrate how exposed people already are to this kind of tracking using entirely off-the-shelf tools, and they said they did not release the underlying code. It remains one of the most concrete illustrations of what happens when consumer face-search tools get combined with wearable cameras and automated lookups.

Opting out and removing yourself

If you find yourself in results on a tool like PimEyes, most of these services offer some form of opt-out or removal request, typically requiring you to verify that you are the person in the photo before they'll act on a removal. It's a reasonable first step, but treat it as damage control rather than a permanent fix: new crawls can pick up new photos, images can resurface on new sites, and removal from one tool doesn't touch any of the others.

A few practical habits help more than any single opt-out request:

• Periodically search your own face on more than one of these tools to see what's actually indexed about you.
• Ask site owners directly to take down old photos you no longer want circulating, since removing the source image is more durable than removing it from one search index.
• Tighten privacy settings on social accounts where your face-tagged photos are public by default.
• Be skeptical of any service that promises to permanently and completely erase your face from "the internet" — no legitimate tool can guarantee that.

Which Ban the Bots page do you actually need?

• Want the legal landscape around facial recognition in the US and UK? Read Facial Recognition Laws.
• New to facial recognition generally and want the basics? Start with our facial recognition explainer.
• Curious which specific companies deploy facial recognition and how? See Company Facial Recognition.
• Want practical steps for protecting your privacy overall? Visit Fighting Back.

Frequently asked questions

Is it illegal to search for someone's face on PimEyes? In most places, using the tool itself isn't illegal, though what you do with the results — stalking, harassment, unauthorized surveillance — very much can be, and some jurisdictions with strict biometric privacy laws have separately challenged how these tools operate.

See Facial Recognition Laws for how that legal patchwork actually breaks down.

Do I need to pay to see full face-search results? Free tiers on tools like PimEyes typically show that a match exists without revealing the full source link, while a paid subscription unlocks complete results — which is precisely the pricing structure that makes them accessible to casual stalkers, not just researchers.

Can these tools identify someone from a blurry or partial photo? Accuracy drops significantly with poor image quality, extreme angles, or partial face visibility, though results still sometimes surface plausible matches worth double-checking rather than trusting outright.

Conclusion

Face-search engines turn the internet's scattered photo archive into a searchable index, and that convenience cuts both ways — useful for spotting stolen photos or catfishing, dangerous in the hands of someone trying to track you down. PimEyes, FaceCheck.id, and Yandex all do some version of the same thing, and the I-XRAY project showed just how far that capability can be pushed when paired with wearable cameras.

If you want to understand your legal footing, read Facial Recognition Laws. For the technology basics, see Facial Recognition. For general deepfake and identity-protection guidance, check How to Spot a Deepfake and Fighting Back.