About
We buy the asset nobody has priced.
Every company that has operated for a decade is sitting on a training-grade corpus. Almost none of them know what it's worth, who buys it, or how to hand it over safely. That's the entire job.
Corpus exists because two things became true at once. Frontier AI labs ran out of useful public text and started paying seriously for real operational data. And nearly every mid-market and enterprise company had a decade of that exact data sitting in Google Workspace, Microsoft 365, Slack, Zoom, HubSpot and Dropbox — costing them storage fees and returning nothing.
So we started buying it. We scope what you have, price it against live demand, run the anonymization review that makes it legally and ethically transferable, and manage the handover into our secure enclave. The corpora we acquire are used to train and evaluate business AI systems — never resold, never passed on.
You are paid on close. You never pay Corpus a fee, a retainer, or a percentage. If your archive turns out not to be worth the effort, we'll tell you that in the valuation rather than run you through a process.
Principles
How we operate.
Price before commitment
You see a written valuation range before signing anything. No discovery process that traps you into a deal.
Anonymization is not optional
We would rather lose an engagement than transfer a corpus that still carries identifiers.
One counterparty
We are not shopping your data across an open market. You license it to us, once, and it goes nowhere else.