Real people, their own devices, the rooms they actually live and work in. Speech is what we have shipped — thousands of hours, nine locales, consent on every file. Video, image and text run on the same recruiting and QA machinery.
Every number here is speech work already shipped. New modalities are listed the same way — after they ship, not before.
One line says delivered because it is. The rest say open because the bench is ready for them and the first orders are being scoped — we will list their hours the same way once they ship.
Two-party calls, spontaneous speech, read prompts, accents to quota, and the same script in a car, a kitchen and a warehouse.
Task video from a phone or an action camera — third-person and first-person — plus the embodied capture on the Physical Intelligence pages.
Photographs to a spec — products, scenes, documents, faces with consent — across the devices and lighting your users actually have.
Written dialogues, prompts and chat transcripts authored by people who work in the sector, in the register the sector actually uses.
Not the dataset — the input your model gets wrong. A clip, a photo, a paragraph is enough to start.
Quantity, locales, devices, conditions, annotation depth — and the contributor rate you will be paying, stated as its own line.
Five percent of the volume, delivered first, so the spec breaks early instead of at the end.
Drops into your bucket with a quality report and the reviewer chain attached to every batch.
Defaults, negotiable in the written spec — but nobody should have to ask what a delivery contains.
Full delivery spec and termsContributors record on their own phones, in their own rooms, in the accent they speak. Blue is where the bench works today; anything else we recruit to order against locale quotas.
Pick a session, talk for twenty minutes on your phone, get paid once it passes review. No studio, no interview, no shift you have to show up for.
The base rate is $20 per approved audio hour. Conversation sessions, noisy-room sessions and locales we are short on pay above it — and whatever a session says it pays is what you are paid, shown before you start.
Send one input your model handles badly — a clip, a photo, a paragraph. We will tell you what would fix it, and what it would cost, before anyone signs anything.