Real-time sign language recognition converts hand gestures to text and captions — so deaf, mute, and hearing people can connect naturally, in any conversation.
Built from the ground up for the deaf and mute community — and everyone who wants to communicate with them.
MediaPipe Hands tracks 21 landmarks per hand. Our gesture engine classifies signs using actual finger angles and landmark geometry — not random guessing.
Detected signs and spoken words appear as animated captions on the other person's screen in real time — no interpreter needed.
Hearing users speak naturally. The Web Speech API transcribes every word as a live caption stream so deaf users can follow every sentence.
Detected signs are read aloud via text-to-speech synthesis, letting mute users have their gestures voiced for hearing participants.
Full-screen remote video, draggable PiP local view, hand skeleton overlay, caption history panel, and one-click controls — all in one.
All messages — text, sign, and voice — are stored locally and synced. Sign messages show their ASL label and auto-generate captions.
From gesture to caption in under 100ms — here's the pipeline.
Your webcam streams video to MediaPipe Hands, which finds your hands in every frame.
21 3D points per hand are tracked — fingertips, knuckles, wrist — 60 times per second.
Our engine calculates finger extension angles and landmark ratios to match your gesture with confidence scoring.
The recognized word appears as a live caption on the remote side — instantly, automatically, no interpreter needed.
Create your free account and start a conversation that everyone can understand.