Free pilot batches for qualified AI teams. See how it works
How it works

See our quality before you sign anything

A proven four-step process, starting with a free pilot batch on your own data.

Four steps from first call to production

  1. 01

    Scoping call

    20 minutes on your model, data type and acceptance criteria. We sign an NDA before you share anything.

  2. 02

    Sample batch

    We agree the batch size together. Send it securely, with your guidelines or without; we can draft them.

  3. 03

    Delivered with a QA report

    Annotated in your format, with accuracy, agreement scores and every flagged edge case.

  4. 04

    Scale when ready

    Project pricing or a dedicated team, run by the same people who did your pilot.

Inside every project

Plan

Guidelines aligned with your model objectives, quality metrics and acceptance criteria defined up front.

Annotate

Domain-trained annotators in Dhaka, with real-time quality monitoring and guideline refinement.

Review

Two-layer QA, edge-case identification and benchmarking against your standards.

Deliver & scale

Integration with your pipeline, weekly reporting and workforce scaling as you grow.

What affects the price

We don't publish a price list, because two projects with the same number of images can need very different effort. These are the things that move the price.

Because delivery is done by our own team in Dhaka, our prices are typically below European and US annotation companies.

  1. Data typeImages, video, medical scans, text, audio or multimodal data. Video and 3D take longer per item than single images.
  2. Label complexityA few bounding boxes are quick; detailed polygons, pixel-level segmentation or long descriptions take more time.
  3. Volume and timelineLarger volumes lower the cost per item. Tight deadlines need a bigger team.
  4. Quality levelSingle review or two-layer QA, and the accuracy and agreement targets you set.
  5. EngagementPer-asset pricing for fixed-scope projects, or a monthly rate for a dedicated team.

Your pilot settles it. The free pilot measures the real effort on your own data, so the proposal you get afterwards is based on facts, not estimates.

Common questions

How big is the free pilot batch?

We agree the batch size on the scoping call, based on your data type and what you need to see to judge our quality.

Who qualifies for a free pilot?

AI teams with a real annotation, validation or evaluation need. Very large or highly specialised pilots may be run as a small paid pilot, credited against your first invoice.

Do you sign an NDA?

Yes, before you send any data. We can also sign a data processing agreement under GDPR.

Which tools do you work in?

Your own tooling or the annotation platform you already use. No tooling yet? We provide our own annotation platform, built on CVAT, customized for our QA process and hosted in Europe, at no extra cost. You get owner access to your project and we deliver in the format your pipeline expects.

Can we see the work while it happens?

Yes. On our platform you get owner access to your project: you can see every label, follow progress and export at any time. Nothing is hidden behind a monthly report.

What happens to our data after the project?

You export your results yourself. If you want, we keep the project for 1 to 3 months in case you need changes; during that time only project owners keep access. When the hold ends or the project closes, we delete the project and its data.

What happens after the pilot?

You get the annotated batch, a QA report and a proposal for the full project, with pricing based on the measured effort.

Start with a free pilot batch

We agree the batch size on a short call. If the quality isn't there, you've lost nothing.