FaceCapture Web brings KBY-AI’s face-recognition technology to the browser: a JavaScript implementation that automatically detects a user’s face and captures a high-quality image of it, with no manual framing or shutter press. The project is published on GitHub, where a live demo lets anyone try the capture flow in a real browser.
The hard part of web onboarding, attendance and access flows is getting a usable face photo — users hold the camera badly, lighting is poor, and blurry shots fail downstream matching. FaceCapture Web handles that capture step in the front end: it watches the camera feed, detects and tracks the face, and automatically grabs a clean image when conditions are right, instead of asking the user to time a button press. It is built on KBY-AI’s Face SDK, whose underlying recognition algorithm the company reports as top-ranked in NIST’s FRVT 1:1 verification evaluations. The repository sits inside KBY-AI’s wider product family — face detection, face recognition, liveness detection and KYC verification SDKs for web, mobile, desktop and server — and the linked demo page shows the same capture experience end users would get. The project was still being updated in 2026.
The online demo is free to try in your browser. The production Face SDK behind it is a commercial KBY-AI product, and licensing is arranged by contacting the company — which also sells related SDKs for recognition, liveness detection and KYC verification.
It is for developers building web onboarding, eKYC, attendance or access-control flows who need reliable automatic face capture in the browser, backed by an enterprise-grade recognition engine.