FieldMesh captures first-person egocentric video inside real US commercial environments — not controlled labs, not staged sets. The actual physical world your models will operate in.
Every environment below is a working US business with employed workers performing real tasks, captured first-person with phone-based headsets workers already own.
Distribution centers, 3PL facilities, e-commerce fulfillment. High-volume, repeatable tasks at scale.
Light and heavy manufacturing lines, assembly stations, quality control benches.
Commercial kitchens, processing plants, co-packers, bakeries, cold storage.
Electricians, plumbers, HVAC techs, welders — complex manual dexterity in variable environments.
Grocery stores, big-box retail, specialty stores. Human-customer interaction mixed with physical tasks.
Non-clinical healthcare environments — hospital supply chains, sterile processing, facilities management.
Commercial cleaning, building maintenance, janitorial operations across multiple site types.
Active construction sites, property inspection, infrastructure work in complex unstructured environments.
Farm operations, nurseries, landscaping — outdoor and semi-outdoor environments with natural variability.
Footage is egocentric first-person, captured via phone-based headset worn by the worker performing the task. What you get is what humans actually see and do — not a top-down camera, not a synthetic render.
Phone-based headset captures exactly what the worker's hands, body, and field of view experience. Native egocentric data — no conversion from overhead cameras.
Not a set, not a stage. Actual operating businesses with normal workers doing normal tasks. Ambient noise, lighting variation, interruptions — reality included.
Task categories, start/end timestamps, environment tags, and worker action annotations delivered with every dataset. Ready for downstream training pipelines.
Every worker opts in individually with full written consent. Commercial usage rights fully cleared. No ambiguous crowd-sourced legal exposure.
Employed workers in stable environments — not gig workers rotating through. Capture recurring task sequences for longitudinal dataset building.
Submit your task definition and camera requirements. We match environments to your protocol, not the other way around. Custom scopes available.
Standard specs for all environment captures. Custom protocols available — submit your requirements and we scope accordingly.
| Parameter | Standard | Notes |
|---|---|---|
| Capture device | Phone-based headset | Workers use devices they already own — zero hardware deployment friction |
| POV | Egocentric / first-person | Head-mounted, captures field of view aligned with worker gaze |
| Environments | US commercial, indoors + outdoors | 50+ active sites across 9 industry sectors |
| Workers | Employed, stable roles | Not gig workers — consistent task performers in recurring environments |
| Consent | Individual written consent | Opt-in only, no penalties for declining, commercial rights cleared |
| Labeling | Task categories + timestamps | Action segments, environment tags, per-clip metadata |
| Delivery | Secure cloud transfer | NDA and DPA executed prior to any data transfer |
| Minimum scope | Custom — contact for scope | Pilot scopes available for new protocol development |
| Turnaround | 24 hrs protocol review | Collection timelines depend on protocol complexity and volume |
Submit your task definition, environment requirements, and volume target. We'll respond within 24 hours with site matches and timeline.
Submit a protocol intakeIf you operate in any of the environments above, your workers may qualify. Submit the form and get a custom revenue proposal in 24 hours.
Get a revenue proposal