Egocentric Video
First-person footage from contributors navigating daily life — markets, transit, workplaces, homes. Urban and rural environments. Multiple countries, demographics, and lighting conditions.
Real humans. Real environments. Verified at scale. TanahLabs sources and delivers first-person behavioural data from Southeast Asia to train the next generation of physical AI.
THE GAP IN EMBODIED AI
Physical AI systems need to understand how humans actually move through space — the texture of a crowded market, the rhythm of a morning commute, the logic of an unfamiliar kitchen. That data does not exist in a lab.
20-70%
drop in task performance when robot models encounter unseen objects, layouts, or viewpoints
$4.4B
market size of embodied AI in 2025
400M+
internet users across Southeast Asia — the vast majority accessing the web mobile-first
DATASET CATALOGUE
Every dataset is collected in real environments, verified through our identity infrastructure, and delivered to your specification. Browse by type below.
First-person footage from contributors navigating daily life — markets, transit, workplaces, homes. Urban and rural environments. Multiple countries, demographics, and lighting conditions.
Ship us your proprietary capture devices, smart glasses, or spatial camera setups. We handle local contributor onboarding, equipment logistics, on-ground operation, and data transfer.
Continuous POV capture in complex, unstructured environments—traditional markets, busy family-run kitchens, manufacturing workshops, and street-front retail.
Bespoke data collection designed around your specific embodied AI use case. Environment specification, demographic requirements, volume, and delivery timeline — all configurable.
Talk to Us →The Pipeline
Every dataset is collected in real environments, verified through our identity infrastructure, and delivered to your specification. Browse by type below.
Verified contributors in our SEA network use their existing smartphones to film first-person footage during daily tasks. Simple app. Clear task brief. Fair compensation paid on delivery.
Every submission passes through our identity and quality layer — contributor verification, frame quality scoring, metadata tagging, and consent logging. Rejection and re-collection handled automatically.
Verified contributors in our SEA network use their existing smartphones to film first-person footage during daily tasks. Simple app. Clear task brief. Fair compensation paid on delivery.
The Study Case
Currently active in 5 key Indonesian cities, including Jakarta and Surabaya. Our verified capture network delivers high-fidelity POV datasets and scales dynamically to any regional hub based on your requests.
Geographic Advantage
Indonesia's combination of urban density, environmental diversity, demographic variety, and mobile-first population makes it the optimal source for robotics training data that actually generalises to the physical world.
4 Active
Collection markets
4 Active
Collection markets
270M+
Population Coverage for Indonesia
20+
Diverse Business Categories
5 Years
Community trust built by W3GG
For Contributors
We don't need you to change what you do. We pay you to share it. Tanah Labs contributors earn supplementary income by filming everyday tasks — from their own point of view, on their own phone.
Trusted Contributor Network
Tanah Labs is a W3GG Group company. We have paid contributors across Southeast Asia since 2021. This is the same community. A new chapter.
For Enterprise
Every dataset delivered by Tanah Labs carries full provenance documentation, verified contributor consent records, and quality assurance against your specification.
Footage is point-of-view recordings of people at work, focused on day-to-day productive tasks. Every contributor consents before filming, and that consent is logged and delivered with each dataset.
Curated collections from specific environment types: dense urban streets, transit systems, traditional markets, residential interiors, outdoor rural settings, and industrial sites.
Curated collections from specific environment types: dense urban streets, transit systems, traditional markets, residential interiors, outdoor rural settings, and industrial sites.
Get Started
Talk to our data team. First engagement scoped within 48 hours.