Data Supplier Program
Supplier programs exist across the data industry. What they actually are, what to check before joining one, and how our own supplier onboarding works. No portal, no registry.
-
A program is a pipeline, not a contract
Joining one gives you a route to work. It does not guarantee volume, and any program that implies otherwise is selling something else.
-
Ask what happens to unsold capacity
The useful question is not how much work exists, but what happens in the months when your language is not in demand.
-
Exclusivity is usually a bad trade
Programs that require it should be able to explain what you get in return. Most cannot.
-
The screening is the value
A program that checks your consent process and quality control is doing work that makes you more sellable elsewhere too.
What a supplier program usually is
In the data industry, a supplier program is a vendor onboarding pipeline. A company that sells data or data services builds a bench of producers it can call when orders arrive, and the program is the process for getting onto that bench. The honest description of what it gives you is a chance at future work, not work itself.
That is not a criticism. A bench is a real asset to both sides. It becomes a problem only when the framing suggests that joining produces income. It does not. Data is bought against a specification, with a defined volume, format, quality bar, and date, so a place on a bench converts to revenue only when a live requirement matches your capability and you deliver against it. And payment follows acceptance: the buyer's QA decides whether a batch is paid, on criteria set before production, not after.
What to check before joining one
Supplier programs differ enormously in how they treat the people on the bench. These are the clauses that decide whether a program is a partnership or a trap.
- Rights assignment: does the program take ownership of your data, or a license, or nothing until a sale? Ownership transfers are the ones to read twice.
- Exclusivity: are you locked out of selling the same capability elsewhere? Some programs require it; most good ones do not.
- Unpaid sample requirements: producing a free dataset "for evaluation" with no buyer attached is work without a customer. A paid or project-attached pilot is a different thing entirely.
- Payment-on-resale models, if the program pays you only when it sells your data, you are carrying the inventory risk. Price accordingly or decline.
- The consent chain, who holds it, who can show it, and what happens to it if the program ends. If you cannot answer that, the buyer's legal review will not either, and an undocumented chain is the most common reason suppliers are turned away in this market.
How our supplier process works
We do not run a registration portal, because we do not buy data on speculation. Suppliers are onboarded against live requirements: when a specification exists that your capability matches, we bring it to you as a project with a buyer, a quality bar, and a payment structure attached. If nothing matches today, your capability goes on file and we say so plainly rather than implying work is coming.
The first message that works is short and factual: languages and dialects you can produce, recording conditions, annotation depth, monthly capacity in hours, and where you are located. Vague introductions stall because there is nothing to match against a requirement. Put that statement in the supplier form at the bottom of this page and it goes on file until a matching requirement appears.
The same honesty applies to what we will not ask for. We do not ask for exclusive lock-in, we do not ask you to assign your data to us, and we do not ask anyone to produce without a defined buyer and a documented consent chain. The declined categories are fixed too: recorded telephone calls, medical or clinical data, and scraped personal data have no path through us, so a capability built on any of those is better taken elsewhere.
What no program can offer you
No supplier program, ours or anyone else's, can give you a route into a major lab through a public submission. OpenAI, Google, and Anthropic do not run upload-and-get-paid portals for third-party data; supply reaches them through data teams, licensing deals, and vendors. Any program that promises a submission channel for a fee is describing something that does not exist.
What a legitimate program can offer is narrower and more useful: a matching function against real specifications, a quality standard that makes your output acceptable when the match happens, and repeat work when you deliver. That is the whole value, and it is enough for suppliers who can produce.
Questions to ask any supplier program
These apply to ours as much as to anyone else's. A program that cannot answer them clearly is either new or not being straight with you.
The answers matter more than the marketing. If a program will not put its terms in writing before you sign, that is the answer.
- What is the payment structure, and what triggers payment? A defined event, or an undefined future one.
- Is there a minimum volume commitment in either direction? If not, the program carries no obligation to send you anything.
- Who owns the relationship with the end buyer? If the program owns it, you cannot follow the work when you leave.
- What happens to data you have already produced if the buyer cancels? Who bears that cost?
- What is required of you beyond production? Consent documentation, metadata format, quality reporting. These are real costs.
- How do you leave? Notice period, outstanding payments, and what happens to data still in production.
- Can you see the specification before quoting? If not, you are pricing blind.
Questions we get asked
Is there a fee to join your supplier list?
No. We do not charge suppliers to be considered, and we would be cautious of any program that does. Our revenue comes from the buyer side of a transaction, not from suppliers.
Will I be told when something matches?
Yes, when a live requirement matches the capability you described. If nothing has matched, we say that rather than sending updates with no substance.
Can individuals join, or only companies?
Capability is what matters. Individuals with rare language access are useful, but most projects require contracts, invoices, and a consent chain, so individual suppliers often partner with a studio or work through one for the paperwork.
Keep reading
-
Sell AI Training Data
If you have recording capability, annotation capacity, or language access that AI teams need, we want to hear from you.
-
Sell Data to AI Companies
How AI companies actually buy data, the routes a supplier can realistically use, and the reasons most first approaches fail before anyone looks at the data.
-
Sell Data to AI Labs
Labs buy differently from product companies: narrower specifications, smaller volumes, and a higher bar on provenance. What a supplier should expect from that process.
-
How we place your data
Who buys the data you supply, and what we handle between the two of you.
Offer your data capacity
Tell us what you can produce, in which languages, and how much per month. If there is a buyer for it, we come back with a real scope, and we carry the contract, the compliance paperwork, and the delivery risk, not you.
- You record; we handle the buyer, the contract, and the paperwork.
- Paid per accepted batch. No listing fee, no exclusivity clause.
- We do not buy medical or clinical data, or intercepted call recordings.
- You tell us your capacity What you can produce, which languages, how much per month.
- We reply within two business days Whether we have a buyer for it, and what that buyer would need.
- You deliver a pilot batch A small run first, so both sides see the quality before a full order.
- We handle the buyer and the paperwork You deliver batches and get paid per accepted batch.