Vision that sees every move
Real-time pose estimation reads the body from a single camera, tracks joint angles, counts reps, scores form quality, and flags unsafe movement frame by frame - no markers, no wearables required.
SELECTED WORK · 06All work
BeBTR - it sees every move.
A preventive-health foresight initiative exploring how AI, computer vision, and wearable data can shift fitness from reactive exercise tracking to proactive injury prevention. It turns any camera into a real-time form coach - on-device, private, and running right in the browser.
- Built for
- Department of Health - Abu Dhabi
- Stack
- COMPUTER VISION · POSE ESTIMATION · ON-DEVICE INFERENCE
1stPlace - Smart Health Hackathon
20k AEDIn award funding (DoH · Abu Dhabi)
30fpsReal-time on-device pose analysis
100%Private - runs entirely in the browser
WHAT IT DOES
Coaching that catches injuries first.
A live form coach that watches your movement, counts your reps, and corrects your posture - all on-device, in real time, before a bad rep becomes a real injury.
Proactive, not reactive
Most fitness tech logs what already happened. BeBTR corrects posture before injuries happen - shifting the model from exercise tracking to genuine injury-prevention foresight.
Private by design
The entire computer-vision pipeline runs on-device, in the browser. No video, image, or body data ever leaves the user's machine - privacy is the architecture, not a setting.
Zero-install, instant demo
Open a link and the AI form coach is live - using the device camera with nothing to download. The same on-device pipeline ships as a real, runnable product, not a slide deck.
RECOGNITION
Built for Department of Health - Abu Dhabi.
BeBTR took 1st place at the Smart Health Hackathon and was awarded 20,000 AED in foresight funding at Abu Dhabi Global Health Week - hosted by the Department of Health and Microsoft. A judged, real-world validation that preventive AI health is worth building.
BUILT WITH
- COMPUTER VISION
- POSE ESTIMATION
- ON-DEVICE INFERENCE
- WEBGL / WASM
- JAVASCRIPT
- WEARABLE DATA
WHERE THIS FITS
The work behind BeBTR
On-device computer vision that turns any camera into a real-time form coach - pose estimation detects unsafe movement and corrects posture, running entirely in the browser.
Built by Team Kanban, the studio's competition and experimental build team, at Smart Health Hackathon 2025 - 1st place, with AED 20,000 in foresight funding. See the full competition record.
Services this build draws on
- Computer Vision & AI SurveillanceCameras already watch your sites - they just can't tell you what they see. We build computer vision systems that detect, monitor, and inspect from existing feeds, with people reviewing what matters so the system stays safe and accountable.Read the service page
- Artificial IntelligenceFrom machine learning and natural language to computer vision and predictive analytics, we build a full suite of AI capabilities - tailored to how your business actually works, with a person reviewing the decisions that matter. No black boxes, no hype.Read the service page
