Consented first-person (POV) video and multimodal data, collected in Latin America, annotated to enterprise standards.
Contributors record genuine daily activities on their own smartphones with a head mount, nothing staged, everything consented.
Guideline-driven: camera setup, head-mount placement, task taxonomies, lighting and diversity specs per project.
Explicit, documented consent captured before recording, critical for egocentric data, auditable per record.
A LATAM contributor network recording in native Spanish and Portuguese, across diverse real environments.
Documented acceptance criteria per batch, multi-tier review before anything is delivered.
Schema-ready outputs with clip-level metadata and a full quality report in your format.
Tracking, temporal segmentation, activity recognition, and action grounding on collected footage.
Real-scene perception data for ADAS, robotics, and physical AI.
First-person context for AI assistants learning from human activity.
Task execution footage for training, simulation, and evaluation pipelines.
Send us a short sample of your capture brief, we'll return a small annotated batch and a fixed-scope quote.
Request a Sample BatchJoin our contributor network in Latin America and contribute to real AI training datasets.
Become a Contributor
Xyntriq