Public sample evidence

Inspect Real Egocentric Data.

Household, workplace and English-narrated previews. Choose a sample and request the current inventory that fits your model.

Household

Mechanical Disassembly and Reassembly

Three reviewed phases from one maintenance recording: release a guarded component, clean the detached blade assembly, and align the guard for reassembly.

Capture
Actor POV
Audio
No narration
Preview
15s, 960 × 540, 24 FPS
Request This Data
Technical notes and evidence

Approved public excerpt. Training access uses a separate buyer license.

Download Evidence

Commercial / workplace

Actor Detail With Wider Work Context.

One actor-aligned POV stream is paired with frontal, side, and whole-body external views of the same work event.

Capture
Actor POV + exo
Audio
Audio removed
Preview
15s, 1920 × 1080, 24 FPS
Request This Data
Technical notes and evidence

Manually aligned within an observed 200 ms tolerance. Observed alignment is not validated sensor synchronization.

Download Evidence

Commercial / workplace

Both Wrists Close to the Work.

Right- and left-wrist ego views preserve near-field manipulation while frontal and wide exo views retain body and workspace context. The workspace exo view is cropped for presentation.

Capture
Dual-wrist ego + exo
Audio
Audio removed
Preview
15s, 1920 × 1080, 24 FPS
Request This Data
Technical notes and evidence

Visually aligned within an estimated 500 ms tolerance; not calibration-grade sensor fusion. Observed alignment is not validated sensor synchronization. The full privacy-redacted view remains in the detailed sample records.

Download Full-View Evidence

Commercial / workplace

English-Narrated Garment Packing

A Philippine contributor describes a complete pack, seal and sort cycle in natural English.

Capture
Actor POV
Audio
Human English narration
Preview
32s, 1280 × 720, 29.97 FPS
Request This Data
Technical notes and evidence

This page proves the visible public sample. It does not imply that unrelated EGXO inventory contains narration, that speech is robot-executable action, or that project-specific timing and rights are identical across releases.

Download Evidence
Explore full sample records, transcripts and QA

Household sample / POV

One Task, Visible From the Actor's View.

The featured excerpt shows task state, hand and object interaction, and reviewed phase progression. The full household page contains all 6 approved public previews.

Inspectable household evidence

Preview the Signal. Then Check the Record.

The public manifest describes all 6 sanitized household previews. The evidence record for this featured clip ties its phase annotations to the published media properties and checksums.

View All Household Samples ↗Download Household Preview Manifest ↗Download Featured Sample Evidence ↗

Reviewed commercial ego and exo evidence

Two Work Events. Two Capture Priorities.

One configuration preserves broader scene coverage. The other moves two ego cameras to the wrists for closer manipulation detail while retaining external context.

Scene-first configuration

Actor Detail With Wider Work Context.

One actor-aligned POV stream is paired with frontal, side, and whole-body external views of the same work event.

Commercial garment inspection and sorting15 seconds, 1920 × 1080, 24 FPS, audio removed

Paired preview: observed alignment only, not validated sensor synchronization.

Technical specifications and QA evidence
Ego
actor POV
Exo
frontal, side, whole-body
Alignment
±200 ms
Privacy review
360 frames

Manually aligned within an observed 200 ms tolerance.

All 360 native-HD master frames and all 360 website-derivative frames were reviewed or mechanically revalidated in order. Tight frame-indexed masks follow the visible head in frontal, side and whole-body views; identifier and reflection masks remain. Audio and source metadata were removed.

Contact-first configuration

Both Wrists Close to the Work.

Right- and left-wrist ego views preserve near-field manipulation while frontal and wide exo views retain body and workspace context.

Commercial garment inspection and handling15 seconds, 1920 × 1080, 24 FPS, audio removed

Paired preview: observed alignment only, not validated sensor synchronization.

Technical specifications and QA evidence
Ego
right-wrist ego, left-wrist ego
Exo
frontal, wide whole-body
Alignment
±500 ms
Privacy review
360 frames

Visually aligned within an estimated 500 ms tolerance; not calibration-grade sensor fusion.

All 360 approved-master frames were reviewed in order. The native-resolution web derivative retained the same tight tracked face mask and fixed identifier redactions, then re-passed full decode, mask-opacity, face-detection and QR checks across all 360 frames. Audio and source metadata were removed.

Environment and capture relationship

Commercial Context. Two Ego and Exo Configurations.

Both samples are commercial data and paired ego and exo evidence. Commercial identifies the workplace environment. The ego and exo relationship identifies how actor-mounted and external views relate to the same visible event.

Review Commercial Inventory ↗Plan Custom Egocentric Data Collection ↗Review Synchronized Capture Requirements ↗Compare Camera Placements ↗

Public sample / narrated workplace task

Human English Narration, Grounded in the Visible Task.

A Philippine contributor performs a garment-packing workflow from a first-person viewpoint and describes the visible steps in natural English. The excerpt preserves a complete pack, seal, and sort cycle.

Complete packing cycle excerptHuman voice + first-person video
Viewpoint
First-person RGB video
Language
Human English narration
Task
Garment packing and sorting
Web derivative
1280 x 720 at 29.97 FPS

This page proves the visible public sample. It does not imply that unrelated EGXO inventory contains narration, that speech is robot-executable action, or that project-specific timing and rights are identical across releases.

Watch the fuller public sample2 min 07 sec
Read the highlight transcript11 utterances
  1. Then we get another shirt,

  2. get another plastic bag.

  3. After getting a plastic bag,

  4. we put in the folded shirt.

  5. After putting in the folded shirt,

  6. we need to make sure that it is tightly

  7. and neatly put inside.

  8. After that, we peel off the tape.

  9. After peeling off the tape,

  10. we seal it and then we finally sort it

  11. with the other completed package shirts.

Open the machine-readable evidence record

Narrated egocentric evidence

Listen to the Signal. Then Inspect the Specification.

The public sample demonstrates one English-narrated workplace task from the contributor's first-person view. The narrated resource explains language modes, required episode fields, VLM and VLA use, and the limits of treating human speech as robot action.

Read the Narrated Egocentric Video Guide ↗Request Narrated Inventory ↗Scope Narrated Collection ↗

Evidence boundary

Public Proof, Precisely Scoped.

These previews prove the visible content and reviewed properties of the published clips. They do not, by themselves, grant model-training rights or establish the volume, calibration, timing, annotations, or rights of a future buyer delivery.

  • Household previews are approved marketing excerpts from first-party collection.
  • The scene-first commercial derivative was manually aligned within an observed 200 ms tolerance.
  • The dual-wrist derivative was visually aligned within an estimated 500 ms tolerance and is not calibration-grade sensor fusion.
  • The narrated preview proves human English speech in the visible sample, not robot-executable action or a timing method that has not been release-verified.
  • Inventory scope, licensing, and delivery are confirmed in writing for each buyer.

Choose the next step

Review Inventory or Scope New Collection.

Tell us which environment, tasks, viewpoints, volume, technical properties, annotations, and rights your model requires.