01
Dishwashing & food-service cleanup
Varied dishware, utensils, sinks, water, residue, sorting, drying.
Street Scout Intelligence
Real-world commerce data for embodied AI
Street Scout captures structured task data from restaurants, hospitality, retail, cleaning, and local service environments — helping AI systems understand how work actually happens in the physical world.
Hero source // pg-scoop-cleaning-preview
Source // Pure Green workflow previewStatus // partner-approved preview
The problem
Lab data is too clean. Commerce is messy.
Robotics systems need to understand clutter, occlusion, human motion, shared workspaces, irregular objects, and real service workflows. Most commercial task data disappears after the work is done. Street Scout turns those moments into structured datasets.
Signal // clutterSignal // shared workspaceSignal // irregular object statesSignal // real service workflow
The task data robots actually need
02
Sweeping, mopping & surface cleaning
Debris, floors, counters, spills, obstacles, repeated cleaning motions.
03
Occlusion & object interaction
Partial views, stacking, hidden tools, hand-object motion, cluttered counters.
04
Crowded-space navigation
People, carts, narrow aisles, moving bodies, social spacing, service flow.
05
Commerce task workflows
Food prep, stocking, checkout, bussing, packaging, delivery handoffs.
Pure Green preview set
Commercial kitchen workflows, not stock footage
PG_01 // SCOOP_CLEANING
01
Sticky scoop cleaning workflow
waterresiduehand-object motionstainless steel glare
gs://street-scout-prod.firebasestorage.app/data_site_media/previews/pure_green/pg-scoop-cleaning-preview/
PG_02 // HIDDEN_OBJECT
02
Object discovery in soapy water
foam occlusionhidden objectsafety contextliquid distortion
gs://street-scout-prod.firebasestorage.app/data_site_media/previews/pure_green/pg-soapy-knife-preview/
PG_03 // LID_WASHING
03
Lid washing and object-state workflow
repetitionsimilar objectstask sequencestate transition
gs://street-scout-prod.firebasestorage.app/data_site_media/previews/pure_green/pg-lids-preview/
These public previews are compressed MP4 outputs from Firebase Storage preview folders. Raw MOV files stay out of the public page and are not loaded in the browser.
AI needs context, not just video.
Raw footage alone is not enough. Useful embodied AI data needs task labels, environment context, object states, action sequences, timing, and outcomes.
task type
objects involved
visible vs occluded
action sequence
environment metadata
workflow stage
completion outcome
edge cases
Data from real service environments
RestaurantsCafésCommercial kitchensRetail shopsHospitality spacesShort-term rentalsCleaning workflowsEvent spaces
Street Scout partners with commerce and service spaces where physical work happens every day.
Trust architecture
Consent-first data collection
Street Scout is not a surveillance product. Data collection is permission-based, environment-aware, and designed for responsible AI development.
Site owner permission required
Worker/participant consent where applicable
No hidden recording
Private areas excluded
Sensitive information can be blurred
No selling personal identity data
Building the real-world data layer for commercial robotics
Street Scout helps AI and robotics teams access structured data from messy, high-variation environments where real commerce happens.