Every reach out,
labelled to the frame.

Real sessions from our own capture, growing every month, delivered raw or annotated per hand to your taxonomy.

Golden-set clip G31: both hands hold an appliance housing while the right fingertips smooth down a black pad at a factory station

Real hand work, captured on factory lines, construction sites and in homes. Labelled per hand, audited in passes, growing every month.

10M+

Scale

curated hours in the Video Datalake by end of 2027

1K+

Diversity

tasks across 20+ industries

Every frame

Quality

labelled per hand, then audited in passes

Where we capture.

Every environment, venue and scene in our capture network. Pick a venue to see what we record there.

APACAfricaEuropeNorth AmericaSouth America

Our rigs and capture protocol, deployed at real work sites across five continents.

A real record from our annotation pipeline.

Clip , the window 00:06–00:38. Click a segment, step frame by frame, or play the timeline.

Real capture

Left hand

Right hand

Left hand

Right hand

Selected segment

Id
Family
Pass
Text

Thousands of judged segments per hour of video.

Our Video Datalake pipeline does the work.

Per hand

Left and right are separate tracks, each judged on its own.

Audited in passes

A later pass can suppress a segment, and has to say why in its own words.

Idle excluded

Gaps are idle time and never count toward accepted hours.

What we can provide with every delivery.

s8-segment.json

{
  "clip": "S8",
  "source": { "view": "egocentric stereo", "fps": 30 },
  "track": "right_hand",
  "label": "Grasp",
  "primitive": "grasp",
  "start": 427.079,
  "end": 430.876,
  "text": "The hand holds the component together with the other hand while using its thumb and index finger to press and manipulate a small piece onto the outer surface.",
  "concurrent": { "left_hand": ["Grasp"] }
}

Start with a free sample.

Pulled from the Video Datalake, raw or annotated. For partnerships, pricing, support or press, use the same form and our team will get back to you.

Thank you for the information

A member from our team will be in touch shortly.

Oops! Something went wrong while submitting the form.

WHO WE WORK WITH

—
Enterprise teams with large-scale video data
—
AI/ML teams building multimodal agents
—
Infrastructure teams deploying on edge or on-prem
—
Product teams adding video intelligence to their platform

Video data infrastructure for physical AI

Datalake

Upload

Developer

Research

Company