Docs

Acceptance

Every clip is graded automatically, once, by the same code, the moment it is submitted. You are charged only for clips that pass. The full criteria, with the exact numbers, are on the acceptance criteria page.

What a clip is scored on

CriterionWeightBar
Clip length30%, blockingAt least the request's minimum length
Resolution20%, blockingShortest side at least 720 pixels
Camera and interaction movement30%Controlled movement measured by the accelerometer
Framing steadiness20%At least 40%
A clip must reach 70% overall, or the higher bar you set on the bounty, whichever is greater. Failing any criterion marked blocking rejects the clip regardless of its overall score.
From Acceptance criteria

Graded once, never by hand

No human reviews a clip to decide whether it is accepted, and we do not re-grade a clip after the fact.
From Acceptance criteria

What an accepted clip does not promise

Read these before you fund a set. They are the limits of what the grader tells you.

The grader verifies that the named object is present and that something physical happened. It does not judge whether the chore was performed well, completely, or the way you would want a robot to perform it.
From Acceptance criteria
A clip is human video with quality metadata, motion samples and a detection trace. It is not a teleoperated demonstration and carries no joint states, gripper signals or per-frame action annotations.
From Acceptance criteria

Distinct homes

Node Data treats one account as one home; we do not verify addresses, and you should read the number as a floor on diversity rather than a certified count.
From Acceptance criteria

Disagree with a grade? Dispute the clip before downloading it, or within 14 days of first downloading it — see buyer terms.

Tell us what your robot gets wrong.

Describe the task. Households film it in real homes, every clip is graded against published criteria, and you pay only for the ones that pass.

Request custom training data

$2.50 per accepted clip · $0 for the rest