Who buys data for AI training, and how a seller gets paid
18 organisations that buy data for AI training, what each takes and how it pays, quoted from their own pages, read 28 October 2026.
Kind of data it takes
Who sells to it
How a seller gets in
When it pays
What it asks for, in its words
Every answer from every organisation
Every row is a single organisation, and every answer expands to the wording it was taken from, the page it came from and when it was read. "Not stated" is used when the page says nothing on the point.
| Organisation | Kind of data it takes | Who sells to it | How a seller gets in | When it pays | What it asks for, in its words |
|---|---|---|---|---|---|
| Adobe Stock | photos and images | not statedNot addressed in Firefly FAQ for contributors | an account, app or dashboard | not statedNot addressed in Firefly FAQ for contributors | not statedNot addressed in Firefly FAQ for contributors |
| Chemin | video footage | a person | a project or job posting | once the data is accepted | Ensure recordings meet the required quality, duration, framing, and environmental… |
| Claru | video footage | a person | a project or job posting | once the data is accepted | Record high-quality egocentric video data that feeds directly into computer vision and… |
| Cloudflare | a website's pages | a publisher or rights holder | an application or enquiry form | as each sale is made | not statedNot addressed in Pay per crawl documentation |
| Created by Humans | books | a publisher or rights holder | an account, app or dashboard | not statedNot addressed in Home page | not statedNot addressed in Home page |
| Handshake | a company's own records | a company | an application or enquiry form | not statedNot addressed in License your company data page | We’re looking for companies with 20+ full time employees & 3+ years in operation |
| Maryah | a company's own records | a company | an application or enquiry form | not statedNot addressed in Business data licensing page | Useful data shows how your business handles a request, makes a decision and gets a result. |
| Mercor | a company's own records | a company | an account, app or dashboard | once the data is accepted | If your company operates in one of these spaces, there's a good chance your workflow data… |
| micro1 | a company's own records | a company | not statedNot addressed in Company data partnerships page | once the data is accepted | We're looking for established companies with 30+ full-time employees, mature business… |
| Mode | a company's own records | a company | an application or enquiry form | not statedNot addressed in Data monetization page | Typical target: 30 full-time desk employees. |
| Murf | voice and audio recordings | a person | not statedNot addressed in Sell voice data page | not statedNot addressed in Sell voice data page | Your compensation depends on the dataset's language, quality, and length. |
| Nyne | a company's own records | a company | an application or enquiry form | not statedNot addressed in Sell company data page | Start with the records your business owns and the rights it can license. |
| Protege | not statedNot addressed in Data providers page | not statedNot addressed in Data providers page | not statedNot addressed in Data providers page | not statedNot addressed in Data providers page | Protege works with data and content providers across domains. |
| Shutterstock | photos and images | a person | an account, app or dashboard | periodically, from a pooled fund | not statedNot addressed in Data licensing and the contributor fund help article |
| Silencio | voice and audio recordings | a person | not statedNot addressed in Home page | not statedNot addressed in Home page | 2.5M contributors in 185+ countries record the languages the open web never captured… |
| TollBit | a website's pages | a publisher or rights holder | an account, app or dashboard | as each sale is made | not statedNot addressed in Bot and agent paywall page |
| Troveo | a company's own records | a company | not statedNot addressed in Business data page | as each sale is made | Your historical data, anonymized to our standard |
| Voices | voice and audio recordings | a person | a project or job posting | not statedNot addressed in Voice data jobs page | Once you sign up, we’ll invite you to projects that fit your language, accent, or vocal… |
Where each answer was read (90)
- Adobe Stock, Firefly FAQ for contributors, read 2026-10-28, Kind of data it takes:
Submit photos Submit vectors Submit videos
- Adobe Stock, Firefly FAQ for contributors, read 2026-10-28, Who sells to it: The page is silent here; it was read in full on 2026-10-28, 3,407 characters long.
- Adobe Stock, Firefly FAQ for contributors, read 2026-10-28, How a seller gets in:
Go to Adobe Stock Contributor Portal
- Adobe Stock, Firefly FAQ for contributors, read 2026-10-28, When it pays: The page is silent here; it was read in full on 2026-10-28, 3,407 characters long.
- Adobe Stock, Firefly FAQ for contributors, read 2026-10-28, What it asks for, in its words: The page is silent here; it was read in full on 2026-10-28, 3,407 characters long.
- Chemin, Egocentric video collection project posting, read 2026-10-28, Kind of data it takes:
you'll record **first-person (egocentric) videos** of your everyday activities using your smartphone
- Chemin, Egocentric video collection project posting, read 2026-10-28, Who sells to it:
We're looking for contributors to join a short-term remote project that helps train next-generation AI vision systems.
- Chemin, Egocentric video collection project posting, read 2026-10-28, How a seller gets in:
We're looking for contributors to join a short-term remote project that helps train next-generation AI vision systems.
- Chemin, Egocentric video collection project posting, read 2026-10-28, When it pays:
**Execution:** **USD 3** per **accepted recording hour**
- Chemin, Egocentric video collection project posting, read 2026-10-28, What it asks for, in its words:
Ensure recordings meet the required quality, duration, framing, and environmental standards.
- Claru, Wearable camera data contributor posting (closed), read 2026-10-22, Kind of data it takes:
Record high-quality egocentric video data that feeds directly into computer vision and robotics research.
- Claru, Wearable camera data contributor posting (closed), read 2026-10-22, Who sells to it:
You will use provided wearable cameras to capture footage of daily activities in your own environment.
- Claru, Wearable camera data contributor posting (closed), read 2026-10-22, How a seller gets in:
This role isn't open right now. Apply generally and we'll match you when projects fit your background.
- Claru, Wearable camera data contributor posting (closed), read 2026-10-22, When it pays:
Payment is per accepted recording hour
- Claru, Wearable camera data contributor posting (closed), read 2026-10-22, What it asks for, in its words:
Record high-quality egocentric video data that feeds directly into computer vision and robotics research.
- Cloudflare, Pay per crawl documentation, read 2026-10-28, Kind of data it takes:
Pay per crawl is a feature of AI Crawl Control that enables site owners to control and monetize AI crawler access to content by setting a price per zone.
- Cloudflare, Pay per crawl documentation, read 2026-10-28, Who sells to it:
Site owners to take control of their content, and charge a fee every time an AI crawler accesses a page in their Cloudflare zone.
- Cloudflare, Pay per crawl documentation, read 2026-10-28, How a seller gets in:
To find out how to join the beta program, reach out to us at Pay per crawl signup
- Cloudflare, Pay per crawl documentation, read 2026-10-28, When it pays:
charge a fee every time an AI crawler accesses a page in their Cloudflare zone
- Cloudflare, Pay per crawl documentation, read 2026-10-28, What it asks for, in its words: The page is silent here; it was read in full on 2026-10-28, 4,818 characters long.
- Created by Humans, Home page, read 2026-10-08, Kind of data it takes:
The AI Rights licensing platform for books
- Created by Humans, Home page, read 2026-10-08, Who sells to it:
Sign up for an account, or designate your agent to take control of your works.
- Created by Humans, Home page, read 2026-10-08, How a seller gets in:
Sign up for an account, or designate your agent to take control of your works.
- Created by Humans, Home page, read 2026-10-08, When it pays: The page is silent here; it was read in full on 2026-10-08, 5,582 characters long.
- Created by Humans, Home page, read 2026-10-08, What it asks for, in its words: The page is silent here; it was read in full on 2026-10-08, 5,582 characters long.
- Handshake, License your company data page, read 2026-10-28, Kind of data it takes:
Business records help AI models learn how companies work.
- Handshake, License your company data page, read 2026-10-28, Who sells to it:
We’re looking for companies with 20+ full time employees & 3+ years in operation
- Handshake, License your company data page, read 2026-10-28, How a seller gets in:
Tell us a little about you and your company to get started.
- Handshake, License your company data page, read 2026-10-28, When it pays: The page is silent here; it was read in full on 2026-10-28, 1,525 characters long.
- Handshake, License your company data page, read 2026-10-28, What it asks for, in its words:
We’re looking for companies with 20+ full time employees & 3+ years in operation
- Maryah, Business data licensing page, read 2026-10-08, Kind of data it takes:
Think email threads, sales histories and solved problems, not a contact list.
- Maryah, Business data licensing page, read 2026-10-08, Who sells to it:
AI companies pay for real examples of how businesses work.
- Maryah, Business data licensing page, read 2026-10-08, How a seller gets in:
Get a detailed assessment → Book a 30-minute call
- Maryah, Business data licensing page, read 2026-10-08, When it pays: The page is silent here; it was read in full on 2026-10-08, 11,956 characters long.
- Maryah, Business data licensing page, read 2026-10-08, What it asks for, in its words:
Useful data shows how your business handles a request, makes a decision and gets a result.
- Mercor, Monetize your data page, read 2026-10-23, Kind of data it takes:
The data your business generates every day is exactly what frontier labs need.
- Mercor, Monetize your data page, read 2026-10-23, Who sells to it:
If your company operates in one of these spaces, there's a good chance your workflow data is exactly what labs need.
- Mercor, Monetize your data page, read 2026-10-23, How a seller gets in:
Choose from 100+ supported tools and link your accounts.
- Mercor, Monetize your data page, read 2026-10-23, When it pays:
Get paid as soon as processing is complete.
- Mercor, Monetize your data page, read 2026-10-23, What it asks for, in its words:
If your company operates in one of these spaces, there's a good chance your workflow data is exactly what labs need.
- micro1, Company data partnerships page, read 2026-10-27, Kind of data it takes:
We license data generated through your company’s everyday operations.
- micro1, Company data partnerships page, read 2026-10-27, Who sells to it:
We're looking for established companies with 30+ full-time employees, mature business operations, and substantial historical operational data.
- micro1, Company data partnerships page, read 2026-10-27, How a seller gets in: The page is silent here; it was read in full on 2026-10-27, 10,621 characters long.
- micro1, Company data partnerships page, read 2026-10-27, When it pays:
We pay in full within 30 days of acceptance
- micro1, Company data partnerships page, read 2026-10-27, What it asks for, in its words:
We're looking for established companies with 30+ full-time employees, mature business operations, and substantial historical operational data.
- Mode, Data monetization page, read 2026-10-27, Kind of data it takes:
Get paid for the operating history your business or nonprofit already owns.
- Mode, Data monetization page, read 2026-10-27, Who sells to it:
Get paid for the operating history your business or nonprofit already owns.
- Mode, Data monetization page, read 2026-10-27, How a seller gets in:
GET A QUOTE FOR YOUR DATA
- Mode, Data monetization page, read 2026-10-27, When it pays: The page is silent here; it was read in full on 2026-10-27, 8,964 characters long.
- Mode, Data monetization page, read 2026-10-27, What it asks for, in its words:
Typical target: 30 full-time desk employees.
- Murf, Sell voice data page, read 2026-10-08, Kind of data it takes:
Voice Data Sourcing - Sell Your Data to Murf AI
- Murf, Sell voice data page, read 2026-10-08, Who sells to it:
After voice verification and finalisation of the contract, we'll process payment directly to your bank account.
- Murf, Sell voice data page, read 2026-10-08, How a seller gets in: The page is silent here; it was read in full on 2026-10-08, 9,503 characters long.
- Murf, Sell voice data page, read 2026-10-08, When it pays: The page is silent here; it was read in full on 2026-10-08, 9,503 characters long.
- Murf, Sell voice data page, read 2026-10-08, What it asks for, in its words:
Your compensation depends on the dataset's language, quality, and length.
- Nyne, Sell company data page, read 2026-10-27, Kind of data it takes:
Nyne pays businesses to license anonymized historical data from the tools they already use: accounting, CRM, scheduling and project management.
- Nyne, Sell company data page, read 2026-10-27, Who sells to it:
Our current program focuses on operating companies
- Nyne, Sell company data page, read 2026-10-27, How a seller gets in:
The first call scopes the project: fifteen minutes to map which systems you have, what that data could be worth, and how the process works.
- Nyne, Sell company data page, read 2026-10-27, When it pays: The page is silent here; it was read in full on 2026-10-27, 5,754 characters long.
- Nyne, Sell company data page, read 2026-10-27, What it asks for, in its words:
Start with the records your business owns and the rights it can license.
- Protege, Data providers page, read 2026-10-08, Kind of data it takes: The page is silent here; it was read in full on 2026-10-08, 21,248 characters long.
- Protege, Data providers page, read 2026-10-08, Who sells to it: The page is silent here; it was read in full on 2026-10-08, 21,248 characters long.
- Protege, Data providers page, read 2026-10-08, How a seller gets in: The page is silent here; it was read in full on 2026-10-08, 21,248 characters long.
- Protege, Data providers page, read 2026-10-08, When it pays: The page is silent here; it was read in full on 2026-10-08, 21,248 characters long.
- Protege, Data providers page, read 2026-10-08, What it asks for, in its words:
Protege works with data and content providers across domains.
- Shutterstock, Data licensing and the contributor fund help article, read 2026-10-08, Kind of data it takes:
Shutterstock datasets are a product offering developed to support companies building computer vision machine learning models.
- Shutterstock, Data licensing and the contributor fund help article, read 2026-10-08, Who sells to it:
The share individual contributors receive will be proportionate to the volume of their content and metadata that is included in the purchased datasets.
- Shutterstock, Data licensing and the contributor fund help article, read 2026-10-08, How a seller gets in:
in January 2023 we added an opt out function in the contributor account settings
- Shutterstock, Data licensing and the contributor fund help article, read 2026-10-08, When it pays:
Compensation: The Contributor Fund
- Shutterstock, Data licensing and the contributor fund help article, read 2026-10-08, What it asks for, in its words: The page is silent here; it was read in full on 2026-10-08, 14,748 characters long.
- Silencio, Home page, read 2026-10-08, Kind of data it takes:
Pre-collected multilingual voice and audio, structured by language, region, and use case.
- Silencio, Home page, read 2026-10-08, Who sells to it:
Every dataset starts with a real person on a real device, recording in their own language, in their own environment, on their own terms.
- Silencio, Home page, read 2026-10-08, How a seller gets in: The page is silent here; it was read in full on 2026-10-08, 7,135 characters long.
- Silencio, Home page, read 2026-10-08, When it pays: The page is silent here; it was read in full on 2026-10-08, 7,135 characters long.
- Silencio, Home page, read 2026-10-08, What it asks for, in its words:
2.5M contributors in 185+ countries record the languages the open web never captured, from Wolof to Quechua.
- TollBit, Bot and agent paywall page, read 2026-10-28, Kind of data it takes:
AI bots and agents access publisher content in fundamentally different ways than humans
- TollBit, Bot and agent paywall page, read 2026-10-28, Who sells to it:
AI bots and agents access publisher content in fundamentally different ways than humans
- TollBit, Bot and agent paywall page, read 2026-10-28, How a seller gets in:
Get started freeRead the docs
- TollBit, Bot and agent paywall page, read 2026-10-28, When it pays:
When an AI system requests content, TollBit evaluates the request, enforces the applicable rate, and processes payment in real time
- TollBit, Bot and agent paywall page, read 2026-10-28, What it asks for, in its words: The page is silent here; it was read in full on 2026-10-28, 3,921 characters long.
- Troveo, Business data page, read 2026-10-27, Kind of data it takes:
AI labs and startups are actively paying six figures for company data exports.
- Troveo, Business data page, read 2026-10-27, Who sells to it:
Turn your company data into recurring revenue.
- Troveo, Business data page, read 2026-10-27, How a seller gets in: The page is silent here; it was read in full on 2026-10-27, 4,973 characters long.
- Troveo, Business data page, read 2026-10-27, When it pays:
Most data buyers pay once and disappear. A Troveo partnership is different: every new product is a new sale that you share in.
- Troveo, Business data page, read 2026-10-27, What it asks for, in its words:
Your historical data, anonymized to our standard
- Voices, Voice data jobs page, read 2026-10-08, Kind of data it takes:
Earn extra cash on your schedule by lending your voice to AI training projects.
- Voices, Voice data jobs page, read 2026-10-08, Who sells to it:
Earn extra cash on your schedule by lending your voice to AI training projects.
- Voices, Voice data jobs page, read 2026-10-08, How a seller gets in:
Once you sign up, we’ll invite you to projects that fit your language, accent, or vocal style.
- Voices, Voice data jobs page, read 2026-10-08, When it pays: The page is silent here; it was read in full on 2026-10-08, 56,262 characters long.
- Voices, Voice data jobs page, read 2026-10-08, What it asks for, in its words:
Once you sign up, we’ll invite you to projects that fit your language, accent, or vocal style.
Method
Each organisation's own the organisations' own data-supply and contributor pages were read through estate-fetch and stored whole (sha256 on every row).
For every page, five questions were answered from its own words: kind of data it takes, who sells to it, how a seller gets in, when it pays, what it asks for, in its words.
Every stated answer carries the sentence it rests on, checked against the stored text; 'not stated' means the page read does not address the question.
A organisation with several pages reads 'varies' where they disagree.
18 organisations, 18 pages.
All 18 sources
- micro1, Company data partnerships (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Handshake, license your company data (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Mode, data monetization (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Nyne, sell company data (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Mercor, monetize your data (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Troveo, business data (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Maryah, business data (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Claru, wearable camera data contributor (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Voices, voice datasets (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Murf, sell voice data (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Silencio (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Shutterstock, data licensing and the contributor fund (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Adobe Stock, Firefly FAQ for contributors (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- TollBit, bot and agent paywall (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Cloudflare, pay per crawl (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Created by Humans (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Protege, data providers (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
- Chemin, egocentric video collection project (from the publisher's own the organisations' own data-supply and contributor pages; reproduced for comment)
Quote this figure or put it on your site
18 organisations that buy data for AI training, what each takes and how it pays, quoted from their own pages, read 28 October 2026.
Kind of data it takes, by organisation: a company's own records 7 of 18, voice and audio recordings 3 of 18, photos and images 2 of 18, video footage 2 of 18, a website's pages 2 of 18, books 1 of 18, not stated 1 of 18 (18 organisations, read 28 October 2026).
Who sells to it, by organisation: a company 7 of 18, a person 6 of 18, a publisher or rights holder 3 of 18, not stated 2 of 18 (18 organisations, read 28 October 2026).
How a seller gets in, by organisation: an account, app or dashboard 5 of 18, an application or enquiry form 5 of 18, not stated 5 of 18, a project or job posting 3 of 18 (18 organisations, read 28 October 2026).
When it pays, by organisation: not stated 10 of 18, once the data is accepted 4 of 18, as each sale is made 3 of 18, periodically, from a pooled fund 1 of 18 (18 organisations, read 28 October 2026).
What it asks for, in its words, by organisation: stated 13 of 18, not stated 5 of 18 (18 organisations, read 28 October 2026).
To cite: "Who buys data for AI training, and how a seller gets paid", updated 2026-10-28, https://howmuchdoailabspayfordata.com/datasets/ai-data-buyers/.
The data as a CSV (90 answers, compiled 2026-10-28)