Get paid to film tasks and speak. License the video and audio that train AI.
A two-sided marketplace for AI training data. Contributors film ordinary tasks from their own point of view and record natural conversations — and get paid for both. AI teams license curated, consent-cleared video and audio for robotics, embodied AI, and speech — with full provenance on every clip.
Your task
Clean your kitchen — filmed from your point of view, 5 to 15 minutes.
The marketplace at a glance
Consent-cleared & licensed
Ways to get paid
Approved earnings ever reversed
Reviewed by a person before sale
How it works
From your phone to a licensed dataset
Three steps for contributors. Review, licensing, and payouts are all handled behind the scenes.
Sign up and pick a task
Create an account, accept the license, and browse the video and voice tasks that are open. About a minute.
Film it — or record it
Do the task the way you normally would, filmed from your own point of view — or take a voice task and record right in your browser. No app to install.
Get paid by the hour
A reviewer approves your clip and your balance grows. Withdraw once it clears.
One marketplace, two sides
Built for the people who record — and the teams who build
For contributors
Turn chores you already do — and your voice — into income. Film a task from your own point of view or record a conversation whenever you like, and watch your balance grow.
- Earn up to the posted hourly rate on every approved clip
- Withdraw via Wise, PayPal, bank transfer, or MoMo
- A person reviews every clip, and you see the rate and the reason for it
Example:Video tasks pay up to $12/hour with English narration — up to $6/hour in any other language.
Available to withdraw
$128.40
Household tasks · POV
Egocentric video · 1080p · MP4
For AI teams
License curated point-of-view task video and natural speech audio to train and evaluate embodied AI and voice models — with the paperwork already done.
- Human-reviewed footage and audio, filtered by task, language, and setting
- Versioned license and full provenance on every clip
- Real homes, real hands, real voices — captured first-person, not staged in a lab
Built to be sellable
Every clip carries its own paper trail
The hard part of training data isn't collecting it — it's proving you're allowed to use it. Consent, review, and a versioned license travel with every clip — video or audio — from the person recording to the dataset.
- 1
Recorded
A consenting contributor films a task from their point of view — or records a conversation.
- 2
Reviewed
A real reviewer watches or listens for quality and privacy.
- 3
Licensed
A versioned license + provenance is attached.
- 4
Delivered
Bundled into a dataset and licensed to a buyer.
Voice and video are biometric data, and we treat them that way — no identifiable bystanders, no screens or documents with personal details, and a reviewer checks every clip before anything is sold.
Built for AI teams
Power anything that listens, speaks, or acts in the world
Now collecting everyday tasks and natural conversations — cooking, cleaning, repairs, real talk
Your everyday tasks — and your voice — are worth something. Start earning from them today.
Join contributors recording the footage and conversations that teach AI how the real world looks, sounds, and works — or license a dataset for your team.