AI coaching,
built for athletes
who demand the best.
Real-time pose analysis, shot recognition, and instant feedback — trusted across yoga and cricket, expanding to every sport.
Coaching 15+ shot types across 2 sports, powered by YOLO-Pose and cloud ML on Azure.
Three systems, working in concert.
From effortless biomechanics to instant feedback — every Gaara AI product is engineered on the same core stack.
YOLO-Pose
Ultralytics pose estimation tracks the batter frame by frame — head, hips, feet, and bat path for movement coaching.
Learn more
PyTorch Transformer
A pose-sequence model classifies 15+ cricket shot types from keypoint trajectories — trained on broadcast and academy footage.
Learn more
Form Scoring
Biomechanical rules score stance, head position, footwork, and follow-through — with progress history per athlete.
Learn moreFrom the team.
Building Coach AI: YOLO-Pose and Transformers on Azure
How we moved from browser-only prototypes to a cloud ML pipeline — YOLO-Pose, PyTorch Transformers, and annotated replays for cricket coaching.
Read postWhy your yoga form score matters more than how long you hold the pose
Most fitness apps reward time. We reward form. Here's why correct biomechanics is the only metric that drives real progress.
Read postFrom research lab to webcam: democratising biomechanics
Pose analysis used to require a $50k motion-capture lab. Now it runs on your phone. Here's how we got here — and what it means for sports.
Read post
Yoga form,
corrected in real time.
5 biomechanical criteria. 10+ poses. Coach-grade feedback from any browser.
Explore Yoga
Batting,
analysed at 30fps.
YOLO-Pose tracking and a PyTorch Transformer classify your shot — with biomechanical coaching once exclusive to professional academies.
Explore CricketSupported by Industry Leaders
Built at IIT Madras and Nirmaan — with cloud credits from AWS Activate, Microsoft for Startups (Azure), and the Y Combinator startup pack.
“Most young athletes never get objective feedback on technique — only whoever is watching that day. Gaara AI turns a phone camera into a consistent coaching layer: measurable movement, clear standards, and progress you can track over weeks, not just one nets session.”
That is the kind of innovation we want to see from the IIT Madras ecosystem — rigorous engineering applied to a problem millions of players face every day, without needing a lab or a full-time coach at every practice.