one API, a lot of doors it opens

Where teams put SightRadar to work.

The same four operations — index, search, compare, detect — power a surprising range of products. Here are the ones teams build most, all on a high-accuracy, Rekognition-compatible API.

Start free — no card See it in action

Event & wedding photo-tagging

Let guests find every photo they appear in from a single selfie. Index thousands of event shots, then search by face — the headline use case for photographers and event platforms.

Index · Search

KYC & identity verification

Match a selfie against an ID document photo during onboarding. Calibrated 1:1 similarity scores you can threshold, dropped into your signup flow in an afternoon.

Compare · Detect

Security & access control

Grant entry by recognizing enrolled employees or residents. Pair detect + search to confirm a face belongs to your authorized collection before unlocking a door or turnstile.

Index · Search

Touchless attendance

Mark check-in and check-out with a face instead of a card or PIN. Accurate, fast, and contactless for offices, campuses, factories, and events.

Marketplace & dating trust

Verify that a profile photo is a real, single, live-looking face — and that the same person isn't running duplicate accounts. Quality-gate uploads before they go live.

Detect · Compare

Financial services

Add a face-verification step to high-value transactions or account recovery. An extra factor that's hard to phish, backed by scores you control.

Compare

Education & campus

Identify students and staff for attendance and access to labs, libraries, and exam halls — with per-tenant isolation so one institution's data is never searchable by another.

Index · Search

Hospitality & contactless check-in

Recognize returning guests for faster, contactless check-in and personalized service, without storing card data or building your own recognition stack.

Photo & content tagging

Auto-suggest who's in a photo across your library or social product. Group faces, dedupe people, and let users tag once and never again.

Index · Search

Retail & loyalty

With opt-in consent, recognize enrolled loyalty members for a personalized in-store or kiosk experience — recommendations, fast checkout, and tailored offers.

Member & visitor management

Recognize returning members at gyms, clubs, and co-working spaces, or pre-enrolled visitors at a front desk — faster entry without badges or sign-in sheets.

Index · Search

Account recovery & anti-fraud

Add a face check to password resets and high-risk actions, and flag when one person is quietly running many accounts. A strong signal that's hard to fake.

Compare · Search

how you would build it

Three of these in detail

The operations, the order to call them in, and the part that will bite you. Each one links to a full build guide.

Event & wedding photo delivery

Who builds this: Photographers, event platforms and venues sitting on thousands of photos nobody can find themselves in.

How it goes together

1Create one collection per event, so galleries never cross-search and deleting an event is one call.
2Batch-index the gallery as it uploads — up to 1,000 photo URLs per call, results delivered to your webhook, at the batch rate.
3Give each guest a selfie box. One search returns the ranked photo IDs they appear in; you render their personal gallery from your own storage.
4Use the reveal tier on each match: auto for high-confidence, review for plausible-but-lower, so a wrong face is never shown as certain.

Batch index · Search · Delete collection

What to watch

The consent story is the product here, not a footnote: guests are searching for themselves, so collect the selfie for that purpose only and delete the event collection on the retention schedule you promise them.

Full build guide: event & wedding photo face-tagging

KYC & selfie identity verification

Who builds this: Fintech, marketplaces and any signup flow that needs the person to match their document.

How it goes together

1Call Compare with the ID document photo and the live selfie. You get one calibrated similarity score back — nothing is stored.
2Pick your own threshold from your own matched and unmatched pairs, then band the result: accept, manual review, reject. Never ship a single hard cut-off you have not measured.
3Run Detect first if you want to reject unusable captures (no face, too small, too blurry) before spending a comparison.
4Log the score and the decision, not the images, so you can audit an outcome without retaining biometrics.

Detect · Compare

What to watch

Face match is one signal, not KYC on its own. It proves the selfie resembles the document; it does not prove the document is genuine or that the person is live. Pair it with document authentication and liveness, and keep a human in the loop for the review band.

Full build guide: add 1:1 face verification to signup

Touchless attendance & check-in

Who builds this: Offices, campuses, events and sites replacing cards, PINs or a paper sheet.

How it goes together

1Enrol each participant once from a consented photo into a collection scoped to that class, shift or event.
2At check-in, search the collection with the camera frame. A match above your threshold is an attendance record; below it, fall back to a manual check rather than refusing entry.
3Store the match result and timestamp in your own system. The collection holds embeddings, not your attendance ledger.
4Delete the collection when the term or event ends — that is the whole retention story, in one call.

Index · Search

What to watch

Attendance is where consent gets skipped most often, and it is the use case most likely to be regulated where you operate. Enrolment must be voluntary with a non-biometric alternative available, and this is not a licence to track movement between check-ins.

Full build guide: face recognition attendance system

Rates and worked examples are on pricing; the endpoints are in the API reference.

Built for consented, lawful use.

Face data is sensitive. SightRadar is for use cases where you have a lawful basis and, where required, consent to process the people involved. We don't support unlawful surveillance or covert tracking — see our Terms and Privacy Policy.

Have a use case in mind?

Spin up a key, top up a few credits, and try it against your own photos today.

Start building — it's free