NetraRT Data annotation, without the cloud
GitHub Download
Annotate Anything
For Anyone

Local, collaborative data annotation for computer vision.
Prompt to segment. Export clean datasets. No cloud.

COMING SOON

From raw pixels to clean datasets

Annotate, refine, and export — without the cloud.

Prompt Segmentation

Click or type, get a pixel-perfect mask.

Bounding Boxes

Draw once, auto-propagate across a clip.

Polygons & Keypoints

Fine-grained labels for anything you see.

One-click Export

COCO, YOLO, Pascal VOC — no glue code.

Annotate for
Segmentation Masks Bounding Boxes Keypoints Polygons Medical Imaging Drone Footage Autonomous Driving Industrial QA Segmentation Masks Bounding Boxes Keypoints Polygons Medical Imaging Drone Footage Autonomous Driving Industrial QA
OUR MISSION

Data annotation, without the tax

As easy as typing. As small as the chip it ships to.

  1. 01 · Why

    Annotation is the bottleneck of computer vision — clunky tools, cloud lock-in, per-seat bills. So teams pay the annotation tax. Born from SAM3.c and Kolosal AI, Netra ends it.

  2. 02 · Today

    A local, collaborative annotation platform. Prompt to segment, refine with a click, export to COCO, YOLO, or Pascal VOC. No cloud. No setup hell.

  3. 03 · Tomorrow

    Your dataset, pruned into a tiny model, shipped to any chip — camera, drone, robot, phone, or Pi. Annotation shouldn't end at export; it should end at deployment.

A teenager films a video, auto-annotates every frame, trains a YOLO, and flashes it onto a chip — on their laptop.
That's today.
EARLY ACCESS · 2026

The first 0 builders
get NetraRT before anyone else.

We're opening NetraRT to a small cohort of computer vision teams tired of paying the annotation tax. Drop your email — we'll send an invite when your slot is ready.

One email. No marketing. Unsubscribe with one click.