First-person POV video
Smartphone-based egocentric footage where appropriate, focused on visible hands, tools, materials, and task progression.
Human-work footage for robotics data teams
Scalable egocentric work data for physical AI
HumanLens helps robotics and physical AI data companies access first-person POV footage from real workforce environments across India, collected under buyer-defined protocols and delivered as raw or lightly QA'd data.
- Smartphone-based POV capture
- Worker and site consent
- Raw or lightly QA'd delivery
Core message
HumanLens helps physical AI data companies access scalable, rights-cleared first-person work footage from real-world environments
What we collect
Action-rich footage of real human work
We are building HumanLens around first-person human-work footage that shows hands, tools, objects, materials, workflows, and task outcomes. The ideal data captures practical environments where people interact with the physical world.
Task and environment metadata
Basic metadata can describe vertical, task type, capture conditions, duration, and collection context.
Buyer-defined protocols
Collection can be structured around your requested verticals, task lists, camera format, metadata, and consent needs.
Raw or lightly QA'd delivery
HumanLens can supply low-processing data for downstream partners to parse, annotate, package, or resell.
Verticals
Practical work environments with visible physical tasks
Pilot collections can be shaped around the data category, location constraints, capture protocol, and usage rights a partner needs to test.
Construction and apartment fit-outs
Tiling, drilling, fitting, flooring, painting, fixture installation, and other hands-on site workflows.
Manufacturing and factories
Assembly, inspection, packaging, tool use, repetitive workflows, and process-oriented shop-floor tasks.
Warehousing and logistics
Picking, packing, sorting, scanning, palletizing, staging, and other movement-heavy workflows.
Nursing-school and training environments
Training simulations, equipment handling, and procedural learning settings. Healthcare-related data is approached conservatively and only in appropriate training or consented environments.
Commercial kitchens and service environments
Prep, cleaning, sorting, tool use, workflow repetition, and other structured service tasks where first-person footage can show object handling and process flow.
How it works
A simple pilot path from requirement to delivery
Buyer defines requirements
You tell us the verticals, tasks, capture format, metadata, and consent requirements.
HumanLens organizes collection
We coordinate workers and sites and collect smartphone-based POV footage where appropriate.
Basic QA and secure storage
We screen for basic usability, consent coverage, hands/tools visibility, and privacy risks.
Delivery for downstream processing
Data can be delivered raw or lightly QA'd so your team can parse, annotate, and package it.
Compliance and data handling
Built around clear consent, privacy screening, and buyer-specific protocols
HumanLens is building its compliance and data handling process around buyer requirements and applicable local rules. Pilot collections are designed to clarify collection scope, permissions, storage, transfer, and usage rights before scaling.
- Worker and site consent
- No audio by default unless specifically requested and consented
- Basic privacy screening
- Secure cloud storage and controlled transfer
- Metadata manifest available
- Buyer-specific collection protocol support
- Intended for AI, robotics, physical AI, and world model data pipelines
Pilot collections available
Testing demand for a specific data category?
We are currently speaking with physical AI data companies, robotics data platforms, and world-model teams about pilot collections. If you have a required capture protocol or target vertical, we can discuss a small sample collection before any larger supply agreement.