PreTrainData helps companies license their data to leading AI innovators. Fair offers, plain-English terms, and you decide exactly what’s shared — and for how long.
We believe companies should be rewarded for the data they create. Everything we do is built around fairness, transparency, and long-term partnerships.
Competitive offers that reflect the real market value of your data — benchmarked, not guessed.
No confusing legal language. A one-page summary sits on top of every agreement.
Choose what to share, how it’s used, and for how long. Exclude anything sensitive.
Help power AI that benefits businesses, industries, and society — on your terms.
Our partners range from 20-person startups to public companies. Here’s what that looks like so far.
AI teams need high-quality, real-world data. If your company has years of it, there may be a buyer — even if it’s never left your servers.
Support tickets, chat logs, call transcripts — de-identified.
Reports, manuals, articles, archives and long-form writing.
Contracts, filings, claims, and structured financial records.
Product photography, inspections, footage, and annotated media.
Repositories, configs, technical Q&A and engineering docs.
Recorded calls, podcasts, voice data and multilingual audio.
Not sure if your data qualifies? Ask us — the assessment is free →
A simple, transparent process so you can explore opportunities with confidence. No commitment until you sign.
Tell us about your data and your goals. A 30-minute call is usually enough.
Day 1We review a sample under NDA and send a clear, competitive offer.
~1 weekYou choose the scope, usage rights, and duration. Negotiate anything.
Your paceYour data powers real AI innovation — and you get paid.
OngoingEvery dataset is different. Here are a few examples of what licensing has looked like in practice.
A regional carrier licensed 12 years of de-identified claims notes to a frontier lab — without sharing a single policyholder record.
A support platform turned 40M anonymised conversations into a one-time payment, with strict usage limits on how the data is used.
A specialist publisher licensed its back catalogue to three AI teams on non-exclusive terms — and kept full ownership of every article.
Licensing data is a big decision. We built every step to protect your customers, your IP, and your reputation.
Can’t find what you’re looking for? Our team is happy to talk it through.
Talk to our team →Most text, image, audio, video, and code data that your company owns or has rights to can be licensed. Common examples include support conversations, documentation, archives, and media libraries. We’ll tell you quickly whether there’s demand.
No. You grant a license with a defined scope, use, and duration. Ownership and intellectual property stay with you.
Personal and sensitive information is excluded or de-identified before any data is shared, and you decide what is in or out of scope. Your privacy and legal teams review everything before signing.
It depends on volume, uniqueness, quality, and current demand. We review a sample and give you a concrete offer — there’s no cost or commitment to find out.
Vetted AI developers building models and products. Your agreement names the permitted uses, and you can exclude buyers or use cases you’re not comfortable with.
We reach out to companies whose data may be valuable for AI training. If you’re not interested, just let us know and we won’t follow up.