
Computer vision & documents
Classification, bounding boxes, segmentation and attributes for images and documents, done by experts from the field. They pass a test on your own images before they label anything paid.
What you get
Export as CSV, JSON or COCO, or pull results through the API and webhooks. Every label carries who made it, when and how long it took.
{
"image": "site_1305940.jpg",
"label": "170302 Bituminous mixtures",
"boxes": [[412, 188, 640, 402]],
"labeled_by": "ex_4b81",
"seconds_spent": 19,
"gold_check": "passed"
}Best for
Computer vision teams in industry, construction, recycling, medicine or quality inspection.
Use cases
Sorting construction waste into categories such as concrete, asphalt or soil, by people who know the difference.
Defect boxes and segmentation on production images, with your class definitions.
Annotation that needs domain knowledge, done in the platform without downloading files.

Who labels
Labelers are recruited for the domain, for example civil engineering or geoscience for construction waste. Each must pass your test, typically 50 images, before paid work.
Quality
Typically 50 items. You set the pass mark (default 90 percent, IoU 0.7 for boxes) or grade the answers yourself. Unpaid for experts, free for you.
A share of production items (default 5 percent) has a known answer. Anyone who falls below your threshold is paused automatically.
Choose a review sample or two to three independent raters on every item. Disagreements go to a review queue.
Who made it, when, how long it took, which test they passed and how it was reviewed. Stored with every label and available to you.
Process
You share sample data, your classes or rubric, and at least one example per class plus rules for borderline cases. We turn it into the briefing every expert reads.
We open the task to the pool for your field, for example civil engineers for construction waste or lawyers for legal evaluation. Experts accept your project terms before they see data.
Each expert works through the same test items, usually 50. Scores are calculated against your reference answers, or you grade the answers yourself.
Compare scores, speed and individual answers in the portal and approve the experts you want. Experts who failed can practice and retry.
Approved experts label or rate your data. Reference items, second reviews and the review queue run in the background.
Download CSV, JSON or COCO from the portal, or receive results through the REST API and webhooks as they come in.
FAQ
You pay the expert time. Example: at 20 seconds per image and €50 per hour, one image costs about €0.28, or €0.33 with a second review on 20 percent of images.
Your images, your classes and one example image per class, plus your rules for borderline cases. We turn them into the briefing the experts see.
No minimum and no setup fee. The first 50 images are a free pilot.
Images are shown inside the platform through signed links and are not offered as downloads.
Related

Training & alignment
Teams fine-tuning or aligning models for law, medicine, finance or other expert domains.

Testing & safety
Teams that need a trustworthy score before a release, a customer pilot or a regulatory review.

Testing & safety
Teams preparing a launch, a safety review or AI Act documentation for a high-risk use case.
Send your items and guidelines. Several experts take the test, you decide who continues.
Request a free pilot