About
Built for people who know things.
We connect contributors of all kinds with the AI labs and research teams building the next generation of frontier models. Some are lawyers, doctors, engineers, and scientists. Others bring native language fluency, cultural knowledge, or everyday device use — exactly what AI development needs. You may be an expert and not realize it.
Why we exist
AI needs human knowledge. At every level.
The bottleneck in frontier AI development isn’t compute — it’s human judgement. A model can’t learn to reason about legal precedent without a lawyer. It can’t understand how people actually search without someone who does it every day. Both matter.
We optimise for quality across the full range — from deep domain specialists to generalists whose language, culture, and lived experience is exactly what AI systems need to improve.
500K+
Contributors in our global network
155+
Locales supported across our network
14
Secure facilities across 3 continents
Our principles
What we stand for.
Selective by design
Our assessment is a genuine filter, not a formality — but it’s designed to find what you’re good at, not rule you out. Generalists and specialists are evaluated on their own terms. This is how we maintain quality and justify paying competitive rates.
Pay reflects expertise
Domain experts earn more because their work is harder to replicate. We don’t flatten rates. A lawyer’s evaluation of a legal brief is worth more than a generalist’s — and we price it that way.
Privacy first
All contributor work is covered by NDA. Client identities and project details are not disclosed. We take confidentiality seriously on both sides of the relationship.
Our commitment
Work that pays fairly. Support when it matters.
We believe the people building AI deserve to be treated well. Here’s what that looks like in practice — based on what we actually do, not what sounds good.
Rates stated clearly upfront
You see what a project pays before you commit to it. No hidden conditions, no surprises after the fact. Premium pay structures apply on safety and content moderation programs where the work is harder.
Consent before sensitive work
If a project involves sensitive or potentially distressing content, you’ll be told before you begin — and your consent is required before you’re assigned to it. You can opt out without penalty.
Support when you need it
All contributors have access to our Employee Assistance Program — covering counseling, legal, and financial support. Regionally adapted wellness resources are available across our global network.
Your knowledge is in demand.
No recruiters. No interviews. Meaningful work and real compensation — on your schedule.