AI Data For Sale

AI training data, licensed from the source

We find the companies that hold the data your models need, check their rights, and set up a license you sign directly with them. If you hold data AI teams want, we find you vetted buyers.

Request training dataList your data

A free 30-minute initial consultation for buyers and sellers. We reply to every enquiry within two working days.

Example data requestIllustration
Need
Customer support calls with transcripts, US English
Volume
5,000 hours minimum
Labels
Speaker turns, resolution outcome, product area
Rights basis
Recording disclosure and privacy notice reviewed by counsel
Privacy
Card numbers muted, names and account IDs replaced
License
Training and evaluation, non-exclusive, 3 years
Delivery
Prepared export, seller to buyer, after signing

Two sides of every deal

We work for both sides of a data license, and we never sit in the middle of the data itself.

For AI teams

Tell us what your model needs to learn. We find holders, vet their rights and arrange samples.

  • Private data you can’t scrape or buy off a shelf
  • Provenance documented for disclosure rules
  • License signed directly with the holder
How buying works

For data holders

Find out what your data is worth and who will pay for it, without sending us a single record.

  • Free assessment from a description, not files
  • Rights, privacy and exclusions worked out first
  • Success fee only when a deal completes
How selling works

How a deal comes together

Consultation

Thirty minutes on what you hold or need. Nothing is shared.

Rights check

Ownership, consent basis, contracts and exclusions, before any sample.

Vetted match

We introduce counterparties we have checked, with your permission.

Direct license

Sample under NDA, terms agreed, license signed between you.

Data AI teams are licensing now

Eight of the 24 data types in our catalog. Each has a specification sheet covering volume, buyers, sources, rights, privacy and published prices.

Conversational dataConversational data is multi-turn dialogue between people, or between people and software, recorded with enough context to show how a conversation unfolds.Language and textSpeech and audio dataSpeech datasets train speech recognition, voice assistants and audio language models, and buyers value accent, language and acoustic variety.Speech and audioVideo training dataVideo model builders license footage they cannot scrape: unpublished, high-resolution and with clear rights.Vision and videoSource code datasetsPrivate codebases with real commit history, tests and reviews are worth more to coding model builders than public code, which most labs already have.Code and agentsAgent trajectory dataAgent trajectories are step-by-step recordings of a person completing a task in software: every click, keystroke, screen and result.Code and agentsEmail archivesBusiness email archives show how real work gets done: requests, negotiations, approvals and follow-ups across long threads.Language and textRLHF and preference dataPreference data records which of two or more model responses a person judged better, and why.Human feedback and evaluationMedical imaging dataMedical imaging datasets train diagnostic and triage models, and labels from qualified clinicians are what make them valuable.Vision and video

See all 24 data types

What the market has paid

There is no price list for AI training data, but published deals give reference points. Every figure links to its source.

All AI data prices

DataReported price
Images (most of a 200 million image archive)Freepik, licensing to two large tech companies, 2024$0.02 to $0.04 per image
ImagesDefined.ai, 2024$1 to $2 per image
Unpublished creator footageAI companies paying creators, per Bloomberg, January 2025$1 to $4 per minute
TextDefined.ai, 2024$0.001 per word
Spirit Airlines data (about 100 million emails, 500 million Teams chats, other records)Alphabet, at bankruptcy auction, August 2026$10 million

Common questions

What is AI Data For Sale?

An independent broker for licensed AI training data. We help AI teams find vetted holders of the data they need, and help companies license the data they hold to vetted AI buyers. We introduce the parties and help structure the deal; the license is signed directly between buyer and seller.

What kinds of data can I sell to AI companies?

Private data that shows how real work is done: conversations, tickets, email, code with history, recordings, video, documents, expert judgments and specialist records. See the data catalog.

How much is my data worth?

It depends on scarcity, quality, structure, volume, exclusivity and rights. Published reference points range from about 2 cents per image to $10 million for a large corporate archive. See how to value a dataset and AI data prices.

Do you have datasets ready to buy?

We do not hold an inventory. We find holders who match your request, check their rights and arrange samples. That is how we source data you cannot buy off a shelf.

What does it cost?

The first consultation is free. Sellers pay a success fee only when a deal completes, agreed in writing before any introduction. See fees.

Do you hold our data?

No. We never take custody of raw data. Samples and transfers move directly between the parties.

All questions and answers

Tell us what you need

Book a free 30-minute consultation. Tell us what data you want to buy or license, and we'll tell you honestly whether we can help and what it involves.

Request training dataList your data
  • We reply to every enquiry within two working days.
  • No upfront fees for sellers
  • We never hold your raw data