Get paid for records your company already keeps.

AI teams need real records of how companies work. We buy your email, chat, CRM and project history, de-identify it, and license it. You deal only with us.

No files or system access needed for an estimate.

NK-53162Example record

Small business email and team chat, 52 firms across 12 industries

sender_role
project architect
sent_at
2024-11-07T15:42:10Z
recipients
[EMAIL_118], [EMAIL_7]
subject
RE: Plan review comments, tenant buildout at [SITE_4]
body
[PERSON_12], the city sent back 4 comments. Three are quick fixes on our side. The fourth asks for a second exit from the training room, because the occupant load on our plan is 54 and anything over 49 needs two. We can add a door to the corridor or cut the room to 49 seats. The door keeps your layout and adds about a week. Can you confirm by Friday so we can resubmit Monday?
folder_path
Inbox/Projects/[PROJECT_19]
One record, de-identified. Placeholders such as [PERSON_12] and [COMPANY_4] stay the same across the whole dataset, so threads still read correctly.

Two ways to work with us

Have company data?

Get paid for it. We buy years of work records from US companies, check them, de-identify them and handle every sale. You choose cash, or a larger amount paid on sale.

Need company data?

License it. De-identified records from US companies across industries, each with a datasheet and a provenance file. Request a sample, or tell us what you need and we collect it.

Know company owners? Earn 10% of our offer to the owner, up to $100K a deal, for each company you introduce.

Become a scout

The everyday record of how work gets done

AI models learn real work from real examples. These are the records we buy.

Email
Threads between staff, clients and vendors.
Team chat
Company channels where decisions get made.
CRM
Accounts, deals, activities and notes.
Support tickets
Questions, replies and resolutions.
Code
Repositories with their history, reviews and issues.
Projects and wiki
Tasks, plans, specs and internal pages.
Field and operations records
Work orders, dispatch, inspections and job costing.
Call transcripts
The text of calls and meetings. We never take the audio.

What qualifies, and what we never take

How selling works

  1. Get an estimate. Answer a few questions about your company and its systems. See a range on the spot.

  2. Get two fixed offers. We review a sample under NDA, then put two offers in writing: paid on sale, or cash.

  3. Export and approve. Your admin runs your systems' own exports with our guides. We check and de-identify the data, and you approve what is released.

  4. Get paid. Cash: most of it within 7 days of our accepting the data. Paid on sale: once buyers have paid us for it.

Get your estimate

From the catalog

De-identified datasets from US companies. Each listing shows its stage, sources, years and volume. Pricing on request.

Datasets
IDDatasetIndustryFirmsYearsSizeStage and licenses
NK-21487NK-21487Commercial general contractor project records and mail, 16 firmsConstructionCompany operations archiveConstruction16 firms2013 to 202619.5M records, 6.6B tokensCollected to orderLicenses: Exclusive, Non-exclusive, First look
NK-17205NK-17205Private SaaS codebase with 186K pull requests and review threadsSoftwareCode and engineeringSoftware1 firm2014 to 2026186K pull requests, 2.6B tokensCollected to orderLicenses: Exclusive, Non-exclusive, First look
NK-26915NK-26915HVAC and refrigeration service tickets with technician findings, 23 firmsField servicesField operationsField services23 firms2018 to 20264.6M records, 1.3B tokensCollected to orderLicenses: Exclusive, Non-exclusive, First look
NK-31748NK-31748CNC job shop quotes, job travelers and customer email, 14 firmsManufacturingCompany operations archiveManufacturing14 firms2015 to 20265.6M records, 1.7B tokensCollected to orderLicenses: Exclusive, Non-exclusive, First look
NK-34529NK-34529Freight brokerage TMS loads with carrier sales and tracking email, 12 firmsLogistics and freightLogistics and supply chainLogistics and freight12 firms2015 to 202618.7M records, 4.9B tokensCollected to orderLicenses: Exclusive, Non-exclusive, First look
NK-41736NK-41736Commercial lines agency submissions and carrier correspondence, 16 agenciesInsuranceCompany operations archiveInsurance16 firms2016 to 202613.3M records, 3.5B tokensCollected to orderLicenses: Exclusive, Non-exclusive, First look

Browse all 55 datasets

How we work

  • You deal only with us. You sign with Nulkan. We license the data to buyers ourselves. Buyers never learn your name.

  • Fixed amounts in writing. Each offer is a fixed dollar amount, set before you sign.

  • Estimates checked against offers. We compare every offer with its estimate and adjust our estimates as offers come in.

  • De-identification we measure. We release a dataset only after our process catches at least 99% of each kind of identifier and the result passes a re-identification test.

  • Isolated processing. Each deal runs in its own locked-down cloud account. Your data never goes to an outside AI service.

  • Raw data deleted on schedule. We delete raw data 60 days after the first buyer accepts the dataset, or when the deal ends.

How we protect data

For AI teams: company data with clean rights

Every company dataset is licensed to us by the company that created it, under a signed agreement. We never take scraped data or lists resold by other brokers.

Each one ships with

  • A datasheet
  • A provenance file
  • A QC report
  • A de-identification report

See what your records are worth.

A few questions, then a range. No files, no system access.

Get your estimate