SeeWhatYouSee
Optional video recording of domestic tasks.
Your perspective. Your control. Your approval before anything is shared.
Ten years ago, nobody believed we’d sleep in strangers’ homes. Airbnb didn’t just build a platform — they built trust, one stay at a time. Video in the home feels like a big ask today. We know. That’s why every safeguard exists before you ever press record.
Scroll to see everything
What This Looks Like
Two capture options — both designed to disappear into your day:
Lightweight chest harness
Uses your phone’s camera. Wear it like a necklace. Weighs less than 100g.
Smart glasses
Ordinary-looking glasses with a small camera. No screen, no display.
“Looks like you’re wearing a necklace with a small camera, or a pair of ordinary glasses. The technology is invisible.”
A Recording Session
Start
Tap to start recording, or say ‘Ambrosia, record.’ One gesture.
Live your life
Cook, clean, teach, care. Whatever you’re doing. The camera sees what you see.
Session ends
Tap to stop, or it ends automatically after your chosen duration.
On-device preview
Review your footage on your phone. Nothing has left your device.
AI blur
Children’s faces, screens, personal documents — all automatically detected and blurred. On your device, before upload.
You decide
Approve what you’re comfortable sharing. Delete anything you’re not. No pressure, no judgment.
Encrypted upload
Approved footage is encrypted and uploaded. Stored securely, anonymised, used only for training data.
What the Camera Sees
Object Recognition
Identifying household items, tools, ingredients in real domestic contexts.
Spatial Awareness
Understanding room layouts, furniture placement, walkable paths.
Task Sequencing
The order of operations in cooking, cleaning, organising — start to finish.
Hand-Object Interactions
How hands grip, pour, stir, fold, wipe — the physics of domestic work.
Navigation Paths
How people move through kitchens, between rooms, around obstacles.
Environmental Context
Lighting, clutter levels, surface types — the real conditions robots will face.
Full HD capture optimised for object recognition and spatial mapping.
Smooth capture for tracking hand movements and task sequences.
H.265 compression keeps files manageable without sacrificing detail.
Privacy Architecture
Video in your home is the highest trust you can place in any company. Here is exactly how we handle it.
On-device processing
Face detection runs entirely on your phone. Raw footage with identifiable faces never leaves your device.
AES-256 encryption
Military-grade encryption for all uploaded footage. In transit and at rest.
Deletion SLA
Request deletion of any footage. Removed from all systems within 72 hours. No copies, no backups retained.
COPPA + GDPR
Compliant with children’s privacy laws globally. We exceed minimum requirements.
Raw footage with identifiable faces never reaches our servers. The blur happens on your device. What we receive is already anonymised.
What We Actually Receive
Your kitchen. Your hands chopping vegetables. Your child at the table. The clock on the wall. A document on the counter. Your face as you look down.
A kitchen. Hands chopping vegetables. A blurred shape at the table. The clock is obscured. The document is redacted. Your face is a blur. Only the domestic task remains — the actions, the objects, the spatial context. Everything personal is gone before it leaves your phone.
Household Onboarding
Both adults can review together. We designed Phase 2 so both adults in the household can review and approve together.
Joint review. Both adults watch the processed footage together before anything is shared.
Partner dashboard. Your partner gets their own login to see what’s been recorded, what’s been shared, and what’s been deleted.
Either adult can pause. One tap from either partner stops all recording. No questions, no delay.
Modesty settings
Configure which rooms, angles, and scenarios are appropriate.
Room exclusion zones
Mark specific rooms as never-record.
Time restrictions
Set recording hours that work for your family.
The Labelling Gets Richer
Now you’re not just labelling abstract data — you’re annotating your own footage. The tasks you recorded become the images you label. Context you already have. Scenes you already understand.
Labelling generic kitchen images. Useful, but you’re working with someone else’s context.
Labelling scenes from your own kitchen. You know what’s in every drawer. You know the workflow. The labels are richer because you were there.
The dual-stream compounding: app data + video footage → richer labels → better tools → better data. Each phase makes the others more valuable.
Compensation
Per hour of approved footage. Earned passively while you go about your day.
Video earnings stack on top of Phase 1 app earnings. Two income streams, one platform.
Combined Phase 1 + Phase 2. More for mothers who record regularly.
Video earnings are additive
Phase 1 + Phase 2 = compounding income
The more you record, the richer your labelling tasks become — and the more you earn from both phases.
Trust is earned, not assumed.
We built the privacy architecture before we built the product.