Technology · Powered by vision

Just video.
That's the whole stack.

HITAI is a real-time computer vision platform purpose-built for combat sports. Pose estimation, action classification, and force inference run end-to-end on video — no athlete-worn hardware, ever.

The pipeline

From pixels to broadcast graphics in under 200ms.

01
Capture
Single 1080p / 30fps feed. SRT / WebRTC.
02
Vision
Multi-fighter detection, tracking, 17-keypoint pose.
03
Inference
Action classification, target zones, force estimation.
04
Output
JSON / WebSocket / OBS / NDI overlays. Live.

No sensors. No wearables. No hardware.

HITAI runs entirely on computer vision. A single camera feed becomes pose estimation, action classification, force inference, and broadcast-ready telemetry — in real time.

Read the technology brief →

INPUT

Single 1080p / 30fps camera feed

LATENCY

< 200ms end-to-end

DETECTION

Strikes, takedowns, clinch, ground

TARGETING

Head · Body · Leg classification

FORCE

Inferred ±3% of contact-sensor reference

OUTPUT

JSON · WebSocket · OBS · NDI overlays

Why vision-only

Sensors don't scale. Cameras already exist.

No athlete impactNothing on the fighter. Zero rules friction.
No retrofitWorks with the cameras already in your venue or gym.
Scales infinitelyOne operator can cover a full event card.
Future-proofModels improve over time, your hardware doesn't change.

Production-ready · Today

Put HITAI in your next event.

Book a 20-minute walkthrough with our team. We'll process a clip from your event, show you the live overlays, and scope the integration.

Request demo →contact@hitai.pro

Just video.That's the whole stack.

From pixels to broadcast graphics in under 200ms.

No sensors. No wearables. No hardware.

Sensors don't scale. Cameras already exist.

Put HITAI in your next event.

Just video.
That's the whole stack.