Anonymous data licensing · $100K+ data estates

Sell your company's data. Not your name.

AI companies are paying for the real business records that never made it onto the internet: email, tickets, code, recordings, files. We strip out every name, find the buyer, and license it without your company ever being identified.

How anonymity works
Your name withheld from buyers Nothing leaves your hands before a contract
$10MGoogle's winning bid for a bankrupt airline's emails, chats and code (Aug 2026)[1]
$7.5MThe losing bid, from an AI data company[1]
$1–4 / minReported rates for licensed video footage[3]
$20M+Paid to data owners by one licensing marketplace[4]
$1B+Reported spend by one AI lab on agent-training environments[5]
What AI companies are buying

The web is used up.
Your archive isn't.

AI models have already read the public web. What they can't get anywhere else is how real work gets done: the decisions, the back-and-forth, and what happened next. That lives in company systems like yours.

Email, chat & documents

Years of how your team actually decided things. AI companies use it to train agents that can do real office work.

Buyers: agent-training teams at AI labs and data vendors

Code with its history

Repositories with every commit, the tickets behind them and the reviews on them. A snapshot alone is worth little; the history is the value.

Buyers: coding-model teams, code-data vendors

Tickets with outcomes

Support cases with the fix that worked, claims with the final decision, quotes with the price that closed. The outcome is what makes it valuable.

Buyers: companies building AI agents and test environments

Recordings & transcripts

Calls, depositions, interviews and dictation, especially when paired with human-checked transcripts.

Buyers: speech and voice-AI companies

Footage & media libraries

Raw and finished video, B-roll, photo archives and broadcast libraries you own outright.

Buyers: video and world-model teams, licensing marketplaces

Research & lab data

Experiment results (including the ones that failed), lab notebooks and technical reports that were never published.

Buyers: AI-for-science and life-science AI teams
How it works

Four steps.
One signature.

You stay in control at every step. Nothing is copied, shared or sold until you've approved exactly what goes and who gets it.

1

Confidential inventory

A call under NDA. You tell us which systems you have and how far back they go. We don't touch or copy anything.

● Under NDA
2

Blind buyer match

Buyers see an anonymous profile, like "40-person translation agency, 12 years, 9 language pairs." Never your name.

● Your name withheld
3

Anonymize & package

People, customers, account numbers and secrets are removed before any buyer sees a sample. You approve every exclusion.

● You approve every exclusion
4

License & get paid

You license the data, you don't hand over the company. Terms spell out use, duration and exclusivity.

● You keep ownership
Anonymity, twice

No one learns who you are. Or who your people are.

"Anonymous" means two things here, and we do both: the buyer doesn't learn it's your company, and the data doesn't reveal your employees or customers.

01 · Your company

A blind listing, not a public one

Buyers evaluate an anonymous profile and a scrubbed sample. Your name is shared only if you decide to share it.

  • Company name, domain and location withheld
  • Buyers sign an NDA before seeing any sample
  • You choose which buyers are allowed to bid
02 · Your people & customers

Identities removed before anything moves

Names, contact details and account numbers are replaced with placeholders, so the data keeps its meaning without identifying anyone.

  • People, emails, phones and addresses replaced with tokens
  • Passwords, API keys and account numbers scanned out
  • Any client, project or date range excluded on request
sample_record.eml 
From: Dana Whitfield[PERSON_1] <dana@harborlinguistics.com[EMAIL_1]>
To: Marco Reyes[PERSON_2], Keller & Price LLP[CLIENT_A]
Re: Certified translation for case 24-1187[CASE_ID]

Marco, the Portuguese deposition translation is done and QA'd.
Two terms were flagged by our legal glossary; notes attached.
Call me at (201) 555-0148[PHONE_1] before Friday. Invoice #88213[INVOICE_ID] goes to Keller & Price[CLIENT_A].
Fictional exampleWhat's kept: the work itself (what was asked, what was done, what happened next)What's removed: who
Who we work with

Built for data worth
$100,000 or more.

Anonymizing a company's archive properly is serious work. We only take it on when we expect the result to sell for at least $100,000, so it's worth your time too.

Minimum data estate
$100K+

The companies that clear this bar usually share a profile: established, 20 or more people at some point, and years of records in systems they can still log into. Industries where AI is changing the work (transcription, translation, publishing, media, software, research, engineering) tend to hold the most valuable archives.

The $100K floor is our qualification threshold, not a price guarantee. Final value depends on buyer demand, volume, and what can be cleared for licensing.

Quick check

Tick what's true. Nothing here is sent anywhere.

Tick the boxes that apply.
FAQ

Questions

Usually, yes, as long as you own the data and your contracts allow it. Part of our review is checking exactly that: who created the data, what your customer contracts say, and whether any of it is regulated (patient, student or tax data, for example). Anything that can't be cleared is excluded. We're not a law firm, and we'll tell you when you need your own attorney to confirm something.

Not unless you choose to tell them. Buyers see an anonymous profile and a scrubbed sample, under NDA. Your company name, domain and location are withheld, and we remove details inside the data that would give you away.

They're replaced with placeholders like [PERSON_1] or [CLIENT_A] before any buyer sees anything. The same goes for email addresses, phone numbers, street addresses, account numbers and passwords. You can also exclude any client, project, mailbox or date range entirely.

In most cases you license it: the buyer gets the right to use a copy for AI training under set terms, and you keep ownership. A non-exclusive license can be sold more than once; an exclusive one usually pays more. For companies that are closing, an outright sale is sometimes simpler.

Doing anonymization and rights clearance properly takes real work on both sides. Below about $100,000, that effort rarely makes sense for you. If your archive isn't there yet, we'll say so on the first call.

Not if the systems still exist. Closing companies are often the best candidates. The urgent part is timing: once accounts are cancelled or servers are wiped, the data is gone. If you're in bankruptcy, any sale goes through the trustee and the court, and we can work with them.

AI developers and the data companies that supply them: teams building AI agents, coding models, speech and video models, and AI for science. Which buyer fits depends on your type of data. You approve every buyer before they see a sample.

From the sale, at a percentage agreed in writing before any work begins. You'll know the exact number up front.

Confidential review

Find out what your archive is worth. Quietly.

Three short questions. No data is shared, and your company name stays private until you decide otherwise.