E²-CAP-VT

A lightweight, vision-first head-mounted EGO video capture device

E²-CAP-VT is the lightweight offering in the E²-CAP-V family, built for vision-led real-world capture — preserving first-person binocular EGO capture while cutting wear and deployment cost, ideal for long-duration, multi-camera, large-scale parallel collection.

SYS / PROD 07

MODEL / CAPVT

REV 2026.08

A lightweight vision-capture option

SEC / 02

UNITS / 03

Lightweight Design

A light kit for long-duration, multi-person parallel capture — lower wear and deployment cost for large-scale, multi-camera field data collection.

First-person Vision Capture

Carries the binocular EGO vision philosophy of the E²-CAP-V family to record first-person footage with full task context.

Same-source Data Pipeline

Shares the capture and upload toolchain with E²-CAP-V — collected data flows straight into the UNIData pipeline with no extra adaptation.

Driving the embodied-AI data flywheel

SEC / 03

ITEMS / 03

01 — Robot Skill CollectionRecords first-person human demonstrations to support imitation learning and skill extraction with real operation footage.

02 — EGO Dataset BuildingSustains large-scale first-person data production with lightweight hardware — feeding VLA models and a prior library of human operations.

03 — Behavior ResearchCaptures real human operations and fine motor processes as the scientific basis for robot cognition and action modeling.

Working with the capture ecosystem

SEC / 04

UNITS / 03

E²-CAP-VThe flagship of the same family — full specs and multi-modal synchronized capture.Full specs on E²-CAP-VVIEW →UNIDataMillion-hour real-world operation datasets and skill assets — where collected data goes.VIEW →TaskHallA collaborative collection platform connecting industry experts with SKILL development.VIEW →

Designed for lightweight vision capture at scale

SYS / ACTIONREV 2026.08