Annotate
Labels built to your schema, checked by a second reviewer.
Image, video, text, audio, point cloud and document annotation, in your tool or ours, with every label logged.
What you get.
Image and video
Boxes, polygons, masks, keypoints, polylines and tracks.
Model label correction
Your model labels first. Reviewers confirm, fix or reject each label, and the corrections come back as training data.
Point clouds
3D cuboids and segmentation, aligned with camera frames.
Text and documents
NER, classification, OCR correction and field extraction.
Audio
Transcription, diarization, intent and sound events.
Content moderation
Policy labels for images, video, text and live streams.
A sample of the work.
This is the interface our reviewers work in, or yours if you prefer. The data shown is illustrative.
Labels
- grasp_target1
- bottle_squeeze1
- clamp1
- can1
Review
QA pass · R07 · 01:42
How quality is held.
- Independent reviewers, never the annotator checking their own work
- Inter annotator agreement of 90% or higher held across classes on construction and geospatial work
- Every label logged with annotator ID, timestamp and correction history
- 100% pre delivery sampling
- Pricing unit
- Per task, per hour, or a dedicated seat
- Platforms
- CVAT, Label Studio, Roboflow, Labelbox, Encord, V7, SuperAnnotate, Supervisely or yours
- Formats
- COCO, YOLO, Pascal VOC, KITTI or your own schema, to S3, GCS, Azure or SFTP
Where we have done it.
Half a million images for autonomous navigation, with every file kept on EU servers
Defect annotation on building imagery in CVAT, with a webhook loop between annotators and reviewers.
50,000 facade annotations a week for building condition assessments
Part identification and damage segmentation for insurance claims AI.
200,000 polygons for automated vehicle damage assessment
Other stages
Send us a batch and your guidelines.
Send a small batch and get it back done, free, usually within 24 to 48 hours. You judge the work before you pay for anything.