Pixel-perfect segmentation & classification
We draw the bounding boxes and the masks. You get clean, instance-level labels — most batches delivered within 24 hours.
Two label types. One flat price.
We specialize — narrowly and on purpose. Instance segmentation and classification, done by trained annotators and quality-checked before it reaches you.
Bounding boxes
Tight, axis-aligned boxes around every instance — the fast, reliable baseline for object detection datasets.
Segmentation masks
Pixel-accurate polygons and masks that trace the true shape of each object — for models that need to know exactly where things are.
Instance classification
Every box and mask carries a class label from your taxonomy — separated per instance, never merged.
Built for training pipelines, delivered fast
- Fast turnaroundMost batches come back within 24 hours — drag the slider above to see raw frames become training-ready.
- Your taxonomy, your formatCOCO, YOLO, Pascal VOC, or a custom schema — exported the way your loader expects.
- Multi-pass QAEvery frame is labeled, then independently reviewed before delivery.
- Consistent at scaleTen frames or ten million — the same guidelines, the same price per frame.
One number. That's the whole pricing page.
No tiers, no per-object upcharge, no annotation-type surcharge. Boxes and masks are included in the same flat rate.
Fixed cost. Bounding boxes and segmentation masks included.
- Every instance boxed and masked
- Per-instance class labels
- Independent QA review pass
- Fast turnaround — most batches within 24 hours
- Export in your format (COCO / YOLO / VOC / custom)
- No minimums, no setup fee
From raw frames to training-ready in four steps.
Send frames
Ship us images or video frames and your class taxonomy.
We label
Trained annotators box, mask, and classify every instance.
We QA
A second pass independently reviews and corrects.
You receive
Clean annotations in your format — most batches back within 24 hours, billed at $0.50/frame.