Logistics & warehousing
Parcels, pallets and paperwork, read right the first time.
Annotation for warehouse cameras, dock and yard monitoring, damaged goods and shipping documents.
From the dock door to the delivery note
Every part of the workflow where we can take work off your team.
Inbound
Warehouse
Outbound
Documents
We work with warehouse CCTV, conveyor video, images and documents.
What we label
Every class is written into a guide with your team before the first batch. Files come back as COCO JSON, YOLO, Pascal VOC, CSV, JSONL or your own schema.
Warehouse video and images
Parcels and pallets
Bounding box and track
Box, envelope, pallet, tote
Labels and codes
Bounding box and OCR
Barcode, QR code, address, tracking number
Damage
Polygon and severity
Crushed, torn, wet, open
Safety
Bounding box and keypoints
Person near forklift, vest, blocked aisle
Documents
Shipping documents
OCR and key value
Consignee, weight, item count, reference
Invoices
OCR and line items
Item, quantity, rate, total
How we test the model
The tests that matter for logistics & warehousing models, and what you get back from each.
Count accuracy
Model counts of parcels and pallets checked against reviewed frames.
You get backCount error by scene and object
Damage recall
Damaged packages collected and reviewed as their own set.
You get backRecall by damage type
Extraction accuracy
Every field on shipping documents compared with the source.
You get backAccuracy and errors by field
Where logistics & warehousing teams start
Problems teams bring us, and the first piece of work we would scope.
Damaged parcels only show up as customer complaints.
What we doDamage labelled by type and severity, so the model learns what to catch at the door.AnnotateLabels are read wrong when they are creased or half hidden.
What we doHard labels collected and transcribed, with every unreadable case logged.AnnotateDocument extraction errors reach billing.
What we doA reviewer checks low confidence fields before they are posted.Live reviewRelated work
Twelve times the volume in five weeks, with quality holding at 95 to 98%
Human in the loop support for visual deviation detection on production lines, with a response time requirement under two minutes.
24/7 live review for production line visual AI, quality above 95% by month two
2,000 robot videos described every week by a team built in seven days
Half a million images for autonomous navigation, with every file kept on EU servers
Client and volumes [REAL DATA: Client name and volumes]
Bring us the logistics & warehousing data your model gets wrong.
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.