Pricing
Measured, not guessed.
WeLabelData prices every annotation project individually: a 30-minute scoping call, a free test batch on your real data, then a precise quote based on measured effort. No price list — because no two datasets cost the same.
How we price a project
Three steps from first contact to a number you can put in a budget.
A 30-minute call
You talk directly with the founder — about your task, data modality, edge cases, quality bar, and timeline. We tell you honestly whether we're the right fit.
A free test batch
We annotate a real slice of your data at no cost — building or refining guidelines, surfacing corner cases, and measuring the actual time each unit takes.
A precise quote
You get a per-unit price and a delivery schedule computed from measured work on your data — not from a rate card or someone else's project.
Why we don't publish a price list
Two annotation projects that look identical on paper can differ in real effort by 5–10×. A street scene with three cars is not a street scene with forty overlapping pedestrians; a clean product photo is not a night-time surveillance frame. Object density, occlusion, class count, guideline maturity, and QA depth all move the number.
A published price list has to handle that spread one of two ways: overcharge the simple projects, or hide a risk buffer inside every quote. We'd rather measure. The test batch tells us — and you — exactly what your data costs to annotate, before any commitment.
What actually determines your price:
- Data modality — image, video, LiDAR/depth, audio, text, or multimodal robot data
- Annotation complexity — bounding boxes vs. polygons vs. tracking with re-identification
- Scene density — objects per frame, occlusion, edge-case rate
- QA depth — single pass vs. multi-pass review, target accuracy
- Guidelines — ready-made vs. built together during the test batch
- Volume and timeline — sustained monthly volume prices differently than a one-off sprint
If you want to understand how the industry prices annotation in general — per-object, per-image, and per-hour models and typical market ranges — we wrote an honest breakdown in our Data Labeling Pricing Guide.
What you get with the quote
After the test batch, you receive:
- A per-unit price for your specific data and task
- A delivery schedule based on measured throughput
- The annotated test batch itself — so you can inspect quality before committing
- Annotation guidelines refined on your actual edge cases
No lock-in: the test batch is free, the quote is yours to compare, and the guidelines stay with you either way.
Pricing FAQ
Why don't you publish prices?
Because annotation effort varies 5–10× between projects that look identical on paper. A price list would either overcharge simple projects or hide a risk buffer in every quote. We annotate a free test batch of your real data, measure the actual effort, and quote from measured work instead.
How is my price determined?
By measurement, not estimation:
- A 30-minute scoping call — task, data, edge cases, quality bar
- A free test batch on a real sample of your data
- A per-unit quote and schedule computed from the measured effort
Is the test batch really free?
Yes. We annotate a representative sample at no cost and return it with the quote — you inspect our quality before committing, we get the data to quote honestly. No obligation to proceed.
How fast do I get a quote?
Most computer-vision projects: a few business days from receiving your sample. Complex multimodal or robotics data can take longer, because guideline development is part of the test batch. You'll get a concrete timeline on the call.
Which pricing model do you use — per hour, per image, per object?
Whichever is most transparent for your data — the test batch tells us which. Our Data Labeling Pricing Guide explains how each model works and when it fits.
Start with the call.
30 minutes, directly with the founder. Bring your data questions.