Try this model on one of your photos to see its result here.
Base model
Finds and outlines 80 everyday kinds of objects, from people and vehicles to animals and household items: a box and a pixel mask for each. This is the model /detect runs when no other model is chosen.
- Tasks
- Object detectionInstance segmentation
- Outputs
- bounding boxes, pixel masks (polygons), class names, confidence scores
- Classes (80)
The 80 COCO classes. Class names in results are always in English.
personbicyclecarmotorcycleairplanebustraintruckboattraffic lightfire hydrantstop signparking meterbenchbirdcatdoghorsesheepcowelephantbearzebragiraffe
Specifications
- Accuracy
Boxes
mAP50-950.711mAP500.866COCO val2017, all 5,000 photos, standard COCO metrics, measured by Comexp on the model as served · measured Nov 10, 2026
Masks
mAP50-950.603mAP500.828COCO val2017, all 5,000 photos, standard COCO metrics, measured by Comexp on the model as served · measured Nov 10, 2026
Boxes come from the detector and masks from the segmentation model, both run on every photo.
- Image requirements
- JPG, PNG or WebP, up to 10 MB, any resolution.
- Video requirements
- Video upload on /detect for signed-in users: MP4, WebM or MOV up to 40 MB, analysed at 5 frames per second. Live streams: push SRT to your personal link or let us pull RTSP; results arrive as an event stream.
- Input / output
- In: one image, multipart/form-data field file. Out: JSON with objects, each with class, confidence, bbox [x1, y1, x2, y2] in pixels and polygon (the mask outline).
- Licence
- Use through the Comexp API under our Terms of service.
- Version
- Updated Nov 6, 2026
- Hardware
- Runs in the Comexp cloud: nothing to install, any device that can send an HTTP request. On your own servers or devices: on request.
- API
curl -X POST https://ml.comexp.net/api/recognise \ -H "Authorization: ApiKey <YOUR_API_KEY>" \ -F "file=@image.jpg"Each call costs one detection from your plan, as on /detect. Your API key is in your account settings.