Sell Data to AI
Home Log in AI Data Value Score Pricing API documentation MCP server for AI assistants 26 US company lists Niche lists
For brokers
The Sell Data to AI Rules How to become an AI data broker Data broker business model Buyer programs compared Qualify a company AI training data companies
For data companies
Firmographic data providers Company data API
Seller guides
How to sell data to AI companies Is it legal? Anonymize before selling Score, then make it safe to sell FAQ and glossary
Score any company Check your own company
Data scouts Data brokers Data companies

AI labs pay for specific company data. Our lists show which companies have it.

If a company has a website, it is in our index: 120 million domains, 99.99%+ of the active internet.

Data buyers want companies that fit their rules. Our lists already have these companies, including from high-value niches like chip design, aerospace and biotech.

Last updated: 8 October 2026

Company-level data only: no contacts and no named people. Scores are estimates from public signals, not valuations, not offers and not a sign that a company wants to sell.

See how it works

Score any company, or get the list of companies like it

Every company on our lists carries its AI Data Value Score and the data it likely holds. Lists cover 26 US industries, and the same lists are available for 30+ countries.

Company lists

26 US industry lists of companies AI buyers want data from

Each list is drawn from our index of 120 million domains, and lists keep growing as more companies pass our checks. On every list, every company domain is verified active and every company is checked for fit with its industry. A verified active domain resolves, is not expired, is not parked and shows no error or placeholder page. Every company carries its AI Data Value Score, the data it likely holds, its history and its activity status, so you start with the companies most likely to hold what buyers ask for.

Each sector page shows a free preview with 5 companies checked for that industry, and those rows can change. Live company counts per sector are on the lists page.

Every list is bought once, as a file. It holds the top 3,000+ by AI Data Value Score, selected from the US domains in that sector of our index: from 11,913 in the smallest sector to 236,277 in the largest. Those 3,000 are the ones most likely to hold what buyers ask for. Pick your lists at checkout, and the checkout works out the price.

Top 3,000+By AI Data Value Score, 200 more added every week, selected from 11,913 to 236,277 US domains per sector in our index
$1991 list
$3792 lists
$5393 lists
$7995 lists
+$199each further list
$3,900all 21 lists

One-time purchase by PayPal or card. The lists page shows each list's live count. Download links are emailed right after payment, valid 30 days with 10 downloads per file. Files are a snapshot with no updates. The monthly plans do not include lists: they score the domains you send, from Basic at $99 a month for 5,000 lookups to Pro at $299 for 25,000 and Scale at $799 for 100,000.

The data AI labs pay most for, identified for every company

Not all company data is worth the same. Our research shows AI labs pay most for the records of real work, held in the systems below. We identify which of these systems each company runs, and it is one of the most important factors in the AI Data Value Score.

Work tickets and internal wikiJira, Confluence
Customer support ticketsZendesk, Intercom
CRM recordsSalesforce, HubSpot
Calls and meetingsZoom, Gong
Contracts and documentsDocuSign, Box
Design filesFigma, Miro
Finance and ERP recordsNetSuite, QuickBooks
Code with historyGitHub, GitLab
Work email and documentsMicrosoft 365, Google Workspace
HR and recruiting recordsWorkday, Greenhouse
Training and course contentLearning platforms, course libraries
Fleet and sensor dataFleet and machine data
Clinical and lab systemsPatient and lab records
Human-era archivesYears of records written before 2023

Every company comes with the systems we identified, shown as tags. Company level only: we show which systems a company runs, never what is inside them.

Custom lists to your ICP

Your ICP is not a sector? We build the list for you

One pass over 120M domains. You name the signals, and every company gets match or no match on each one, with the quote and the page behind every match. We narrow it to the IAB verticals of your ICP, keep the live companies, then read each one against your criteria. A ready list stops at a sector; a buyer's rules rarely do: a size band, a system the company runs, a client portal, a state, an ownership type. A company makes your list only when its own pages show each must-have, with the quote and the page it came from.

Two example workflows: 26 US industries available the same way
120M domainsOur full database, 99.99%+ of the active internet, classified. The starting list.
93,020 US domains in IAB Legal Services: narrowed to the IAB verticals of your ICP
Live companiesverified active and checked for fit with the industry
ICP screeningevery candidate screened against your criteria
Your qualified listranked, each signal with its quote and source page; the pilot measures the count for your ICP
120M domainsOur full database, 99.99%+ of the active internet, classified. The starting list.
160,054 US domains in IAB Advertising and Marketing: narrowed to the IAB verticals of your ICP
Live companiesverified active and checked for fit with the industry
ICP screeningevery candidate screened against your criteria
Your qualified listranked, each signal with its quote and source page; the pilot measures the count for your ICP
  1. 1
    Describe your ICPThe kind of company, sectors, geography and size, then your must-have signals, nice-to-have signals and exclusions. Write it the way you would brief a researcher.
  2. 2
    Start the pilot for $100Pay by PayPal or card. We check your profile before the run starts and write to you if anything is unclear. The $100 is credited in full to a full run.
  3. 3
    Get your first 25 companies within 1 business dayEach one with its AI Data Value Score, every signal, the verbatim quote and the source page, plus a projection of how many qualified companies the full run will find.
  4. 4
    Choose the scopeEvery qualified company, a ranked top slice or a fixed number. Full runs start at $999 and are quoted after the pilot, strongest fits first.
Example, fictional companyPilot delivery, row 1 of 25
1
Harbor Ledger CPAsharbor-ledger.example · tax, audit and advisory for closely held businesses
74Score · B
ActiveUnited States22 years onlineFounded 2004
ShownOffers a client portalLog in to the Client Portal to upload your tax documents securelyharbor-ledger.example/clients/ · must-have
ShownTax or audit services for businessesaudit, review and tax planning for closely held businesses across the regionharbor-ledger.example/services/ · must-have
Not shownSeveral office locationsnice to have: reported, never used to drop a company
Example, fictional company: the firm, domain, quotes and values are invented for illustration, on a reserved .example domain that does not exist.
$100pilot, credited in full to a full run
25qualified companies, every value with a quote
1 daypilot results within 1 business day
$999full runs from, quoted after the pilot

Company level only: no contacts, no named people, and we never contact the companies on your list.

Who it is for

For the people who find companies, not the companies themselves

Three groups do the same job from different sides. Each one needs companies that fit a buyer's published rules, and each one loses weeks on companies that never could.

Brokers and referrers

You introduce companies and get paid when they sign

  • micro1"earn $50,000" per company
  • Mode"$50K per referral"

Those are the amounts each program publishes, and they are paid only when a referred company signs and meets the program's conditions. A referral business lives or dies on the companies you bring in. Most companies you know will miss a rule on size, location or records.

The AI Data Value Score shows which companies likely hold the records buyers want. The free check shows whether a target meets the published rules. You approach fewer companies, and better ones.

As published, checked 7 October 2026. Each program's own terms decide who is paid, how much and when.

Data companies

Your partnership team needs supply that meets your own rules

micro1, Mode, Troveo, Miro Advisory and others buy company records and turn them into data for AI labs. Each one publishes who it wants, then has to find enough companies that fit. That search is the slow part of every sourcing plan.

  • Rank your own target domains by score before outreach starts
  • Score inbound applications before your team spends time on them
  • Start from a one-time file of the top 3,000+ scored companies in any of 21 US sectors
  • Send your own domain lists to the API for scoring, up to 100 domains per call

The plans differ only by lookups a month. Sector lists are sold separately, as one-time files.

Context: archive buyers

Some buyers look for companies that are closing

Archives of Slack, email and code can change hands when a company winds down. Troveo cites about $5,000 per code repository and roughly $10K to $100K per archive deal in the shut-down startup market. Forbes (16 April 2026), Fast Company and Gizmodo covered startups selling their old Slack and email.

Our Shutdown Radar lists US companies showing signals of winding down, each with the evidence and the data systems it runs or last ran. Every AI Data Value Score also shows a company's activity status: active, winding down or acquired, parked, or unreachable. If you work this side of the market, that status is on every company you score.

Figures as cited by Troveo and as reported. Context only.

The problem

Buyers publish their rules. Finding companies that meet them is the work.

Every program that pays for company records says who it wants. The rules are short, specific and strict, and most companies fail at least one of them.

A broker's real cost is not the introduction. It is the hours spent on companies that were never going to qualify: a 12-person agency put forward to a program that asks for 30 or more, a strong team based outside the US sent to a program that counts US office staff, a two-year-old startup offered to a buyer that wants several years of records. Each of those looks promising until someone reads the rules.

The rules also differ by program and by industry. Mode drops its minimum to 10 people for accounting firms and 6 for law firms. micro1 asks for 30 or more and puts the US first. A company that misses one program can fit another, so you check every target against every set of rules, not only the one you know best. The table puts them side by side, with where each rule shows up in our tools.

Rulemicro1 (as published)Mode (as published)Where you see it here
Company size30+ employees; the referral posting says 30 to 20020+ full-time US office employeesThe check reads team size from public pages, with a quote
Industry exceptionsNone listedAccounting firms 10+; law firms 6+The score's industry value group; the check's industry reading
LocationUS firstUS office employeesThe check reads location and US presence
LanguagePrimarily EnglishNot listedThe check notes the language of the company's pages
History and recordsMature operationsSeveral years of recordsThe score's history and "online since" year; the check's founding year
SystemsModern software toolsNot listedThe score's likely data held; the check's systems named on public pages
Company payout"$100K-$2M+ for approved data packages""$100K-$5M"Ranges, not a price for any company
As published, checked 7 October 2026, on each program's own pagesFull side by side: buyer programs compared
Two hypothetical targets. A 15-person accounting firm in Ohio clears Mode's 10+ for accounting firms and misses micro1's 30+. A 40-person software company in Texas clears both on size, and its fit then turns on records and systems. Neither is a real company. The point is that a few public facts decide where an introduction can go, and you can read most of them before your first message.

From the whole web to a short list you can work

120 million domainsOur index: 99.99% of the active internet, with the history of each domain.
One sector listFor example, US accounting firms or law firms: verified companies, scored and ranked.
Active, with history and likely recordsSet aside parked and unreachable domains. Keep long histories and the data buyers name.
Meets a buyer's published rulesSize, location, history and systems, against micro1 and Mode.
Your introductionThrough your own channels first, then through the buyer's own program.
What you get

Score, list, connect and qualify

Four products and one free tool, all built on the same index of 120 million domains. Each one answers a question you would otherwise answer by hand, company by company.

Every plan, free demo

AI Data Value Score

Scores any company from 0 to 100 for the data AI buyers want, with a grade, the data the company likely holds, its history and an activity status: active, winding down or acquired, parked, or unreachable.

Nine factor groups go into it: history, scale, knowledge assets, operational systems, customer systems, organization, industry value, expertise and activity status.

Score a company in the demo
One-time files, from $199

Company lists

21 US sector lists, each with verified companies ranked by AI Data Value Score. Each company carries its score, grade, likely data, history and activity status. Buy a list once as a file, from $199 a list (the 3,000+ highest-scored companies), at the list checkout. Each file is a snapshot with no updates. See all 26 lists.

A free preview is public, with 5 companies checked for that industry, so you can see the shape of the list before you pay for anything.

See the free preview
Every plan

API

REST and JSON, with your key in an X-API-Key header. Score a domain from your own CRM, spreadsheet script or enrichment flow, and keep the results next to your own notes on each company.

Score your own domain lists in batches of up to 100 domains per call, one lookup per domain, on every plan. A usage call shows the lookups used and left this month. The API scores the domains you send; it does not return company lists.

Read the API documentation
$100 pilot

Custom lists

When your ideal customer is not one of the 26 industries, we build the list to your own rules: a size band, a system the company runs, a state or an ownership type. Every company comes ranked, strongest fits first.

Start with a $100 pilot for the first 25 qualified companies, credited in full to a full run.

How custom lists work
Free, about 20 seconds

Qualify a target against buyer rules

Enter a company's website. The check reads its public pages and compares team size, location, history and systems with the published rules of micro1 and Mode, with a quote for each fact it finds.

Where the pages are silent, it says "not stated" instead of guessing.

Qualify a target
Everything

Company level only

Every product works at the level of the company: its domain, score, grade, likely data, history and status. There are no contacts, no named people, no email addresses and no phone numbers anywhere.

You reach companies through your own network and your own channels, and you decide what to say to them.

How brokers work
What a score is, and what it is not. An AI Data Value Score is an estimate from public signals. It is not a valuation, not an offer and not proof that a company wants to sell anything. Use it to decide which companies to look at first. Only a buyer can price data, after it sees a manifest and samples, and only the company can decide to sell.
Example result

What a score looks like, and how a broker reads it

This card is a fictional example: an invented lab company on a reserved .example domain, laid out with the same kind of fields a real result carries for any domain you score. To see real companies for the industry, open the live labs and CROs preview.

Fictional exampleAI Data Value Score
78of 100
northfield-labs.example
Life sciences, pharma and labs
Grade BActive
Status
Active
Online since
1993
Company
Invented for this example, not taken from any list

Data this company likely holds

Support tickets and chat transcripts Knowledge base, SOPs and documentation Customer accounts and portal activity Order and transaction records

Fictional company and values. The .example domain is reserved and does not exist. A real score is an estimate from public signals, not a valuation or an offer, and not a sign that the company wants to sell data.

  • Score and grade78 out of 100 is a grade B. On a list, the score is a ranking tool: it tells you which companies to look at first, not what any company's records are worth.
  • StatusActive means the company is operating. Programs for running companies, such as micro1 and Mode, are the natural route for an active company. A different status points to a different conversation.
  • Online since 1993A long web history. Mode asks for several years of records and micro1 for mature operations. A long history is a reason to look closer, not proof that old records still exist.
  • Likely dataSupport tickets, a knowledge base and SOPs, customer accounts and order records: the kind of operational records buyers name. These are likely holdings, read from public signals, not an inventory.
  • What the card cannot tell youHeadcount, who to talk to, or whether the company would ever sell. For size and location, run the check against buyer rules. For a way in, use your own network.
A broker's workflow

From a segment to a closed referral, in seven steps

The first four steps happen here, before you contact anyone. The last three happen in your own network and inside the buyer's program.

1

Choose a segment Here

Start where buyers already want data and where you have reach. Pick one of the 26 US industry lists and look at its free preview first. For an industry outside the 20, start from the companies you already know and score them one by one.

2

Score every candidate Here

Run each company through the AI Data Value Score, in the demo or through the API. Sort by score and grade, and read the likely data held. A company whose likely records match what a buyer names goes to the top of your list.

3

Check activity status Here

Set aside parked and unreachable domains. Winding down or acquired is a different conversation, and often a different kind of buyer. Active companies go forward to programs built for running businesses.

4

Qualify against buyer rules Here, free

Run the check on each shortlisted company: about 20 seconds, with a quote for every fact found. "Not stated" is common for headcount, so treat it as a question for the company, not a no.

5

Approach the company through your own channels You

We provide no contacts. Use your own network, clients and introductions. Say plainly who you are and how you are paid, and never present yourself as a buyer's partner or agent: micro1's terms forbid it.

6

Introduce through the buyer's program Buyer

Once the company agrees, submit it through the program's own referral route. Read the terms before you do. micro1, for example, decides in its sole discretion, pays after onboarding plus a minimum revenue threshold, can claw back payouts, and forbids sharing a payout with the company. Referral programs compared.

7

Wait for the close Buyer and company

Deals take 60 to 90 days to close, practitioners say. The buyer runs the review, the contract, the export and the payment. Keep scoring and qualifying in the meantime: a referral pipeline needs far more names than any single deal.

For data companies

For partnership and sourcing teams: rank the pipeline before anyone reaches out

A data company that buys company records lives on supply. Its own rules decide who qualifies, and its sales targets decide how many qualifying companies it needs each quarter. The gap between those two numbers is filled by outreach, referrals and inbound applications, and every one of those sources brings in companies that were never going to fit.

The market is crowded at the top. A market map by Deedy Das (July 2026) counts 50+ companies selling data and RL environments to labs, with about $8.5B in revenue, 75% of it held by Scale, Surge, Mercor and Handshake. For every other company in that map, finding supply faster is a large part of how it competes.

Where scores help

Prioritizing outreach. Score a segment and start at the top. A company with a high score, an active status, a long history and likely support, knowledge or operational records is a better first message than a random name from a directory.

Screening inbound. When a company applies through your form, a lookup returns its likely data, history and status before your team spends time on it. Parked and unreachable domains show up immediately.

Scoring your own lists in your own tools. Send your prospect file to the batch endpoint, up to 100 domains per call with one lookup per domain, on every plan, and keep the results next to your own notes in your CRM. To start from a ready list instead, buy the top 3,000+ companies of a US sector once, as a file, at the list checkout.

Practitioners say raw data is the cheapest form, evaluations are worth about 10x raw, and full environments reach 6 to 8 figures. Whatever a data company builds on top, it starts with companies that hold the records.

How a sourcing team uses it, week by week

  • Rank a segment by score and grade before the first email goes out.
  • Filter by status: active companies for running-company programs; the rest set aside.
  • Match likely data to what your customers ask for, such as support tickets or SOPs.
  • Screen inbound applications with one lookup each through the API.
  • Batch score your own domain lists, up to 100 domains per call, on every plan.
  • Track lookups used and left this month with one usage call.
  • Start from a sector file: the top 3,000+ companies of a US sector, bought once.
Choose Scale
Plans

Three plans. Pick by lookups.

Every plan uses the same index and the same score, and batch scoring of your own domain lists is on every plan. The only difference is how many lookups you get each month. One lookup scores one domain.

Basic

$99 a month
5,000 lookups a month

For a solo broker scoring the companies in their own network.

  • AI Data Value Score for any company
  • Grade, likely data held, history and activity status
  • REST API with JSON responses
  • Batch scoring, up to 100 domains per call
Choose Basic
For small teams

Pro

$299 a month
25,000 lookups a month

For brokers and small teams scoring several prospect files a month through the API.

  • Everything in Basic
  • Five times the lookups of Basic
  • Batch scoring of your own domain lists
Choose Pro

Scale

$799 a month
100,000 lookups a month

For data companies' partnership and sourcing teams.

  • Everything in Pro
  • Four times the lookups of Pro
  • The lowest price per lookup of the three plans
Choose Scale

No free plan; the demo is free. So is the check against buyer rules. Pay by PayPal or card. Your API key appears in your dashboard after payment, not by email. The plans do not include company lists. Need a ready list? Buy it once as a file, from $199 a list (the 3,000+ highest-scored companies). Full details on the pricing page.
Last updated: 8 October 2026.

Where the companies go

The buyer programs your introductions feed

Amounts and rules are quoted as each program publishes them, as published, checked 7 October 2026. They are ranges and referral amounts, not offers, and the top of a range is not the typical result.

micro1 Enterprise Data Partnership

"$100K-$2M+ for approved data packages"
Who fits
30+ employees (the referral posting says 30 to 200), US first, primarily English.
Referral
"earn $50,000" per company, "no cap", paid after onboarding plus a minimum revenue threshold.
Terms to read
Sole discretion, clawbacks, no sharing payouts with companies, no posing as micro1's partner. The micro1 partnership explained.

Mode

"$100K-$5M"
Who fits
20+ full-time US office employees; accounting firms 10+; law firms 6+; several years of records.
Referral
"Earn $50K per referral". A post on X states up to $55k or 6%.
Terms to read
Mode's own referral terms. Treat its site and terms as the reference, not a post.

Miro Advisory

$100K-$1M+ datasets; $10K-$1M+ code
What it publishes
Indicative ranges: operating datasets $100K-$1M+, codebases $10K-$1M+.
For brokers
A route for software companies with private codebases, as well as operating businesses.
Terms to read
Miro Advisory's own site.
$10MGoogle's agreed price for bankrupt Spirit Airlines' internal dataReported, August 2026
$12.5Mmicro1's competing bid for the same dataApproval not confirmed as of 7 October 2026
~$8.5Brevenue of 50+ companies selling data and RL environments to labsDeedy Das market map, July 2026
60 to 90 daysfor a data deal to closeWhat practitioners say

Spirit was a large airline in bankruptcy, not a typical referral. Use these figures to understand the market, not to promise a company a number. Compare every program on one page in buyer programs compared, and the referral side in data referral programs.

Guides

Guides for brokers and data companies

Each guide quotes buyers only from their own published pages, with the date we checked them.

Company lists

Segments where buyers already look.

Company data

Where company facts come from.

The market

Who buys, and how they work.

Brokering

Referring companies, step by step.

What your target companies will ask

Guides written for the companies themselves. Send them along when the questions start.

FAQ

Questions from brokers and data companies

Do you provide contacts?

No. Everything we sell is company-level: the domain, the AI Data Value Score, the grade, the data the company likely holds, its history and its activity status. There are no named people, no email addresses and no phone numbers, in the demo, the API or the lists. You approach a company through your own network and your own channels.

Is there a free trial?

No. There is no free plan, no free key and no trial. Two things are free: the AI Data Value Score demo, which scores a limited number of companies a day, and the check against buyer rules, which qualifies a target in about 20 seconds. Paid plans start at $99 a month for 5,000 lookups.

How fresh is the data?

Scores are built on our index of 120 million domains, 99.99% of the active internet, with domain history. Every score includes the company’s activity status: active, winding down or acquired, parked, or unreachable. The 26 industry lists keep growing as more companies pass our checks, so we do not print list sizes here. The lists page at https://www.selldatatoai.com/lists/ shows the live count for each list.

Can I resell lists?

Our terms decide what you may do with list data, and nothing on this page grants a right to resell it. Read the terms at https://www.selldatatoai.com/terms/ before you pass list data to a client or anyone else, and treat anything they do not cover as not granted.

Do you sell a shut-down list?

Yes: the Shutdown Radar (https://www.selldatatoai.com/shutdown-radar/) lists US companies showing signals of winding down, each with the evidence and the data systems it runs or last ran. A signal is not proof of closure, and we never contact the companies. Every AI Data Value Score also shows the company’s activity status, so for any company you score you can see whether it is active, winding down or acquired, parked, or unreachable.

Does a high score mean the company wants to sell its data?

No. An AI Data Value Score is an estimate from public signals. It is not a valuation, not an offer and not proof that a company wants to sell anything. It tells you which companies likely hold the records buyers want, so you know where to spend your outreach first. Only the company can decide to sell, and only a buyer can price the data after it sees a manifest and samples.

How do I qualify a target against a buyer’s rules?

Use the free check at https://www.selldatatoai.com/check-your-company/. Enter the company’s website, and in about 20 seconds it reads the company’s public pages and compares team size, location, history and systems with the published rules of micro1 and Mode, with a quote for each fact it found. Where the pages say nothing, the result says so instead of guessing.

Do the monthly plans include company lists?

No. The plans score the domains you send. Basic costs $99 a month for 5,000 lookups, Pro $299 a month for 25,000 lookups and Scale $799 a month for 100,000 lookups, and batch scoring of your own domain lists, up to 100 domains per call with one lookup per domain, is on every plan. Company lists are sold separately, once, as files at https://www.selldatatoai.com/lists/buy/. Each list is the top 3,000+ companies of a US sector by AI Data Value Score, at $199 for 1 list, $379 for 2, $539 for 3, $799 for 5, $199 for each further list and $3,900 for all 21. Download links are emailed right after payment, valid 30 days with 10 downloads per file, and each file is a snapshot with no updates.

Start with the companies buyers want

Score a company in the free demo, qualify it against buyer rules in about 20 seconds, and pick a plan when you are ready to score your own prospect lists.