The public internet has been read. Your archive hasn't.
Every major model has trained on the same public web. Labs now compete on data nobody else has: specialized, accurate, and produced by people who knew what they were doing.
Public data
Scraped years ago and used by every lab.
Forum threads, news archives, image boards, encyclopedias
Your data
Never seen by a model, and built by people with real expertise.
Specialist archives, recordings, field notes, years of professional records
What counts as data
If it was made by skilled people, covers real work in detail, and isn't already on the public web, there's a good chance a lab wants it. You don't need to know which category it fits. Tell us what you have and we'll assess it.
Written archives
Documents built up over years of work.
Reports, correspondence, manuscripts, notebooks
Audio and video
Recordings of real people doing skilled work, in any field.
Meetings, lectures, demonstrations, calls
Records and logs
Structured data from specialized systems and fieldwork.
Logs, forms, measurements, inspection records
Images and scans
Visual material that's labeled, specialized, or hard to find anywhere else.
Scanned documents, diagrams, photographs
Expert corrections
Work where a skilled person checked, corrected, or labeled something.
Edits, reviews, annotations, ratings
How it works
Nothing is shared until you approve it, and buyers only see redacted samples before a purchase.
-
Tell us what you have
Answer a few questions about your data so we can judge what it's likely worth.
About a minute -
Agree on scope
On a short call we look at samples together and agree on what's included and what stays out.
One 20-minute call -
Transfer securely
You send the approved data over an encrypted connection. We clean it and remove personal details.
We do the heavy lifting -
Get paid
Once the data passes review and a buyer licenses it, the payment goes to you.
Paid after review
Tell us about your data
A few questions so we can tell you whether labs are likely to want it and roughly what it's worth. Your answers stay private, and we won't add you to a newsletter.
We review your answers and email you with an initial view on value and next steps. There's no obligation, and nothing is shared until you agree to it.
How we protect your data
Before anything is shared, we agree on these in writing.
A clear scope
- You approve exactly what's included before anything is transferred.
- Original data is kept only for processing, then deleted.
Privacy of the people in your data
- Names, contact details, and identifying information are removed before licensing.
- Nothing is published or shared publicly.
Secure transfer
- Data is encrypted in transit and at rest.
- Access to your systems, if any, ends when the project ends.
Terms you sign off on
- Use, retention, and payment are set out in a signed agreement.
- Your data is never sold to your competitors.
Questions
Who buys the data?
AI labs and research teams training models. They license data to improve how models handle specialized work. Your data isn't published.
What kind of data qualifies?
Specialized material that isn't on the public web and was produced carefully by people who know their field. Larger archives and longer history usually mean more value. If you're not sure, tell us what you have and we'll take a look.
Do I need to own the data?
You need the right to license it. If ownership is shared or unclear, we'll work through it with you before anything is transferred.
What happens to personal information in the data?
Names and identifying details are removed during processing, and the original files are deleted once processing is done.
How much work is this for me?
Usually one call and a secure transfer. We handle cleanup, formatting, and de-identification.
How is the payout determined?
It depends on the type of data, how much there is, how many years it covers, its quality, and the license terms. We'll give you a range after a first look.
When do I get paid?
After your data passes review and is licensed. We'll give you a timeline on the first call, before you share anything.