The annotation platform for computer-vision teams. Draw bounding boxes, polygons, and keypoints, review at scale, and export to COCO & YOLO — without the spreadsheet chaos.
From first upload to production export, every step lives in one fast, keyboard-driven workspace.
Pixel-precise tools with snapping, magnetic edges, and sub-pixel nudging. Switch tools without leaving the keyboard.
Pre-label with your own model or our foundation detector, then correct instead of starting from scratch. 60% faster.
Assign, review, and approve in stages. Catch conflicts before they reach training, with full annotation history.
COCO, YOLO, Pascal VOC, CreateML — one click, versioned and reproducible. Pipe straight into your training jobs via the API.
Every action has a shortcut. A command palette gets you anywhere in two keystrokes.
Drag in images or sync from S3 / GCS. Auto-dedup, group, and split into train / val.
Pre-label with AI, correct on the canvas, then route to reviewers with one click.
Export to COCO or YOLO, or pull a versioned dataset straight from the API.
Start free today — bring your first 5,000 images on us.