EmpathAR uses your camera to detect posture, expression, and movement in real time, translating them into a live Social Battery overlay for everyone in frame.
Best experience on Chrome or Android, which enables focus lock for smoother tracking.
Most of social intelligence is invisible. We read signals unconsciously: crossed arms, a fading smile, a jaw that won't stop moving. But we rarely act on them in time. We miss the moment someone checks out. We talk past the point they've already left the conversation.
EmpathAR makes those signals visible, in real time, as an AR overlay rendered directly over the people in your camera view. It doesn't tell you what to feel. It tells you what's already happening.
The app uses Google MediaPipe, an on-device machine learning framework, to run two models simultaneously: a pose landmarker and a face landmarker. No data leaves your device.
These three signals are weighted and blended into a single 0-100% Social Battery score using an asymmetric smoothing algorithm. Charge is gradual, drain is immediate, mirroring how social energy actually works.
The battery maps to five named states, each with a recommended action.
Each person detected is assigned a persistent label (P1, P2, P3) that survives them leaving and re-entering the frame. The system identifies returning individuals by sampling the average colour of their upper torso (shirt) and comparing it to stored signatures. The count only ever goes up: new shirt, new number.
Multiple people can be tracked simultaneously, each with their own independent battery, state history, and contextual tips delivered as toast notifications.
EmpathAR runs entirely in your browser using WebGL and WebAssembly. No video, no images, no biometric data is ever transmitted to a server. The camera feed is processed locally in real time and immediately discarded, frame by frame. There is no storage, no logging, and no account required. The only data that exists is what you see on screen.
For the smoothest tracking experience, use Chrome. Android devices support focus lock, which reduces landmark jitter during real-time detection.