The market for
AI training data.

Sell the records your company already has. Or train on verified private data nobody else has seen, and test samples in seconds.

Get started
  • Free to list
  • Every company verified
  • Providers and labs stay anonymous

Paste one prompt. Your agent does the rest.

Selling or buying, there is nothing to set up. Copy the prompt into Claude, ChatGPT or Cursor and say what you want.

  1. Copy

    One prompt. It tells your agent what Mora is and how to connect.

  2. Paste

    Into Claude or ChatGPT. It installs the Mora connection by itself.

  3. Sign in

    Once, in your browser. New here? Same page creates your account.

See what it can do

Works inClaudeChatGPTCursorand any MCP client

You have data
You need data
Your agentMora connected

pastedSet me up on Mora (https://mora.market)...

MoramoraMora connected. You signed in with Google in your browser.

What do we have that an AI lab would pay for?

Moralist_connectors

  • Zendeskfound: ZENDESK_SUBDOMAIN in .env.example
  • Slackfound: @slack/bolt in package.json
  • Jirafound: PLAT-2291 in your commits
Claude

I read this repository: you use Zendesk, Slack and Jira. Your Zendesk holds what labs ask for most: tickets from the first message to the fix. Which Zendesk should I connect?

The support one, acme-support.

Moraconnect_tooltool: "zendesk", workspace: "acme-support"

over to youOne click: you authorize Zendesk in your browser. Your agent never sees a password.

Authorize Zendesk
Mora checks it before any lab sees it
  • Mora draws a random sample: nobody picks the best ticketsnext
  • Mora checks no personal data is leftnext
  • Mora checks it is not already publicnext

Example sessions with an agent connected to Mora. The listings shown are previews.

How data gets from a company to a model

Everything runs through your agent. Nothing to configure on either side.

Connect the tools your company already uses.

Years of work already sit in them. That is what labs license.

Chat, email and files
  • Slack
  • Teams
  • Gmail
  • Outlook
  • Dropbox
  • Google Drive
  • SharePoint
  • Notion
  • OneDrive
  • Confluence
  • Google Sheets
  • Airtable
  • Gong
  • Box
  • Discord
  • Zoom
CRM and support
  • HubSpot
  • Salesforce
  • Zendesk
  • Intercom
  • Freshdesk
  • Front
  • Pipedrive
  • Shopify
  • ServiceNow
Engineering, design and finance
  • Jira
  • Linear
  • GitHub
  • Figma
  • QuickBooks
  • Asana
  • ClickUp
  • Trello
  • monday.com
  • PagerDuty
  • GitLab
  • Bitbucket
  • Sentry
  • Xero
  • Stripe
List your data

One click per tool, read-only. Or let your agent find them: it reads your project and asks which workspace to connect.

  1. Connect or upload

    A provider connects a workspace or uploads a dataset: tickets, threads, recordings, logs. Listing is free.

  2. Agents write the schema

    The provider's agent describes what goes in and what comes out. The lab's agent asks in the same format. Nobody fills in a form.

  3. Mora checks it

    Mora draws a sample at random, so nobody can show only the best records. It looks for copies on Hugging Face and the open web, and for personal data.

  4. Labs test, then license

    A lab's agent tries the sample in seconds. The lab signs a license on Mora and the provider gets paid.

Mora does not label data. Labels come from the provider, who knows the work. Mora makes sure what a lab pays for is real, private and not already online.

On the market now

Datasets for sale, checked by Mora. Providers and labs are never named.

DatasetProviderSizeContributorsPrice
Aviation assistant monorepo: Express API, Next.js dashboard and SwiftUI iOS app, with design reviewProvider 733427221 records (commit history 2, design and its review 219)1Open to offers a lab offers a price per record
Air traffic control radio audioPreviewProvider 7B209,400 hours1$10 per hour
Commercial HVAC service callsPreviewProvider 181C412,000 work orders1$0.44 per work order

This market already exists. It has been private.

These are the deals that were reported. Most never are.

  • $14.3BMeta's investment for 49% of Scale AI, a training data supplier. 2025
  • $10BValuation of Mercor, which supplies expert data to AI labs. 2025
  • $250MNews Corp's five-year content license to OpenAI, as reported. 2024
  • $60MWhat Google pays Reddit each year to train on its posts. 2024

Those deals went to companies with lawyers and a broker on call. Mora gives every company with useful records the same access, without the broker.

Private by default

Companies are licensing business records. Labs don't advertise what they train on. Mora protects both. See exactly who sees what

what a lab sees

Provider 181C

company
status
Verified company
listing
Commercial HVAC service calls
mora, in the middle
  • Verifies both companies
  • Draws the sample at random
  • Checks rights, personal data, public copies
  • Signs with each side, names stay hidden
what a provider sees

Lab 03C9

company
status
Verified company
offered
$150k, non-exclusive, 2 years
  • Nobody is named

    Listings never show the provider. Requests never show the lab. Not before a deal, and not after: each side signs with Mora.

  • Checked, not just listed

    Mora checks every company, and every listing for rights, personal data and whether it is already public.

  • Your data stays yours

    Mora reads it to check it and draw a sample. A lab gets the full dataset only after it signs a license.

  • Free to list

    No listing fee and no subscription. You pay nothing until a lab licenses your data.

We've been on the provider side. Mora is built by the team behind PilotGPT, a speech model trained on air traffic control radio.