6 November 2026

Private sale

Private company data, licensed by the companies that own it

Archives, code and recordings, each from one company working with Rightmark. Owners are named at introduction.

18 lots

Company records

Operating records: mail, chat, documents, tickets, CRM and outreach

Lot SE-A

Exclusive: a B2B sales-software company's own records, 7 tools, more than 10 years, up to 15,900 call hours

One company's operating record across seven tools: about 300 to 350 mailboxes, team chat from 2015 to October 2026, a wiki, Jira, support conversations, the CRM and up to about 30,000 recorded sales and customer-success calls. It shows how an established B2B software company sells, supports its customers and plans its work, in private records that span more than ten years and have never been licensed.

years of records: mail, files and team chat from 2015 to October 2026
more than10
tools from one company: mail and documents, chat, a wiki, work tracking, support, CRM, a call recorder
7
mailboxes over the years
about300 to 350
hours of recorded sales and customer-success calls, by the owner's count; may be lower
up to15,900
people over the years; 50 to 120 at a time, about 75 today
about300

What is inside

  • Snapshot: B2B sales-software company · about 12 years old · 50 to 120 staff over the years, about 75 today · incorporated in the US · most staff work outside the US.
  • Email and documents since 2015: about 300 to 350 mailboxes over the years. The admin console's storage page shows 3.29 TB used in all on 29 October, photos included; by the owner's split, files are about 2.6 TB and mail about 372 GB. The owner puts email at about 99% English, mostly with customers. The all-time message count is not known yet.
  • Team chat in Slack, May 2015 to 1 October 2026: about 2.18 million messages in 987 channels by the owner's count. Slack's own analytics, which reach back about a year, count 588,398 messages from members and 82,549 files uploaded from 28 September 2025 to 28 October 2026. The plan and which channels the delivered export covers are being confirmed. The owner expects 15 to 20% of the messages in English. Private channels may be added only with their members' consent.
  • On the tools' own screens on 29 October: 55,028 Jira issues across all projects (20 projects since 2016, by the owner's count); 406,110 closed Intercom support conversations since September 2015; and the CRM's deal records since 2017, 16,306 deals, with the people in them masked. By the owner's figures of 26 October, not yet measured: about 15,200 Notion pages since 2019.
  • Up to about 30,000 recorded sales and customer-success calls from the company's own teams, up to about 15,900 hours, about 32 minutes a call on those counts. The owner's figure is a bulk count and may be lower; the exact count is being measured.
  • Options the owner can add, by his counts of 28 October, not yet measured: the company's accounting records in QuickBooks, about 345,000 transactions since 2016 in one company file and about 13,800 since December 2023 in a second (expenses, sales receipts, deposits, refunds, transfers and journal entries), with customer names and details masked and bank and card numbers left out; and its test-management record in TestRail, an estimated 8,000 to 15,000 test cases and 50,000 to 90,000 test results across 6 projects (490 MB on its own storage page on 29 October: database 233 MB, attachments 257 MB), linked to the issue tracker. Each joins either license, priced once measured.
  • Never licensed, sold or promised to anyone. Some of the company's software vendors keep a non-exclusive right to use its data to improve their own services, and one keeps anonymized data after the contract ends; the owner states the opt-out status per tool before signing. Open to an exclusive license or to non-exclusive licenses.
  • Storage, Jira, support, the CRM's deals and TestRail are shown on the tools' own admin screens (29 October), available to buyers under NDA with names blanked; the wiki, the all-time chat count and the calls are the owner's figures. The calls' years and the English share of every tool but chat and email are still being gathered.

Best for

  • Training and evaluating enterprise agents that work across email, chat, wikis, tickets and the CRM of one real company.
  • Sales and customer-success agents: real B2B sales and success calls with the same company's CRM and support desk.
  • Long-horizon evaluation held out from public data: how projects, customer issues and decisions unfold over years.
  • Seed data for realistic workplace environments: chat, ticketing, wiki and CRM replicas for agent training.
  • Call understanding: summarizing sales and success calls and pulling out next steps, for recognition and understanding only.

Before delivery

  • The owner's rights and a privacy review, tool by tool, before delivery; the owner notifies staff first; staff give written consent to licensing and to transfer abroad where their country's law requires it, EU staff get notice and a window to object, and the personal data of anyone who did not consent, former staff included, is masked or left out. Records of US residents follow the state privacy route counsel sets (California first); people who opted out of sale or sharing stay out.
  • Personal data of people outside the company (prospects, customers, people on calls) is removed or de-identified to the buyer's standard and the law's before delivery; customer data found in internal channels is screened out, and information customers shared with the company in confidence (account details, screens and files shown on calls, support conversations) is screened out of email, chat, support and calls before delivery.
  • Email is screened before any sample: HR, legal, finance and privileged mailboxes out; automated notifications from other tools out; customers' account details, prices, contract terms and pasted customer data removed; names, addresses, phone numbers and signatures of everyone outside the company masked; attachments out unless screened one by one; a human-checked sample per batch, with the result in the manifest.
  • Calls are licensed for recognition, understanding and evaluation; no voice cloning or speech synthesis, and no speaker identification. The consent position of the other party, and whether the recording notice they heard covers licensing for AI training, is checked for each call set. Calls go as audio and transcripts only, with screen-share video left out; calls without proof of a recording notice stay out; participants in the EU, the UK and Canada stay out until counsel sets a route; customers' confidential information is screened from transcripts; the owner exports the calls, and buyers get no access to the call-recording account.
  • Counts per tool and the exact call hours are measured and listed in a manifest before signing; the call price applies to the hours delivered.
  • Samples only with the owner's agreement, as metadata or masked output, for evaluation only and deleted after the agreed window.
  • Onward licensing terms are agreed with the owner, who may bar named competitors.
  • Once you accept the NDA in Rightmark's buyer portal, Rightmark shares the detailed source list, with staff locations by country. The owner's name and product names are shared only when the owner agrees to an introduction.

Left out

  • Always out: health data, data about under-18s, and payment-card and bank-account numbers.
  • Private channels and direct messages in team chat, unless their members consent and the owner agrees.
  • CRM contact and company records (1,222,346 contacts, 500,488 companies) as records, enrichment fields supplied by the CRM vendor, and any bought or scraped contact lists.
  • Support conversations of users whose accounts were deleted or who asked for deletion; customer content shared in support chats; answers written by the support desk's AI agent.
  • Text generated by the tools' own AI features (wiki AI pages, support AI-agent replies, call-note summaries), flagged and left out.
  • Replies and mail the company's product read through a connected mailbox, where they appear in the company's own records.
  • Data the company's customers keep in its product stays out.
  • Source code is not part of this lot.
  • Material that belongs to clients or partners. Other exclusions are being agreed with the owner.

Lot HRT-ARCHIVE

Exclusive license available: more than ten years of operating records from a B2B software company of a few hundred staff

The internal record of a B2B software company over more than ten years, to a few hundred staff today: team chat since 2015, email and documents, more than 72,000 support tickets, about 590,000 CRM records and its work tracking, all in English. It shows how a real company builds, sells and supports software across seven systems, in records that cannot be collected from the public web.

of records; team chat for nearly all of them
12 years
staff today (owner's figure)
a few hundred
team chat messages in the last 13 months alone
about4.09M
support tickets
72,000+
CRM records
about590,000

What is inside

  • Seven systems: email, documents and calendar in one workspace (about 3.5 TB of mail and files, about 100,000 calendar records), team chat (about 60 GB), a support desk (more than 72,000 tickets), a CRM (about 590,000 records), work tracking (about 5 GB), error monitoring for 12 projects (recent months only) and design files (about 5 GB).
  • Team chat runs from 2015 to 2026, with about 4.09 million messages in the last 13 months alone; the team's daily stand-up updates, collected by a bot, sit in the chat too. The owner describes its records as chat-first.
  • A few hundred staff today, at a company that builds software products for business customers.
  • Most of its staff are in India; the owner estimates about 70% of the records come from staff in India and 30% from the US. All records are in English.
  • Delivered de-identified: names and contact details are removed or masked before anything is shared. Support tickets and CRM records pass a privacy review first. The owner reports that staff were notified.
  • The company's own records only: its customers' product data and file storage, HR and payroll, a chat workspace with outside members, third-party prospect data and AI chat logs stay out.
  • The company's code is offered as its own lot (266 repositories with full git history) and can be quoted together with this archive. Links between tickets and commits have not been checked yet.
  • Figures come from the owner's data inventory. A measured manifest per system comes before any contract.

Best for

  • Workplace agents that act across email, chat, documents, calendar, CRM, support desk and work tracking, trained or evaluated on one company's real records over 12 years
  • Customer-support agents: more than 72,000 B2B support tickets with the product and engineering context around them
  • Sales and account-management agents: about 590,000 CRM records from a B2B software seller
  • Company replicas and RL environments seeded with a real software company's chat, documents and tickets
  • Held-out evaluation sets from private records that are not on the public web

Before delivery

  • A measured manifest per system (volumes, first and last years, mailbox count) before contract; figures today are the owner's counts.
  • Rights: the records of every group company in scope are covered, with the signer named; the owner's "no vendor" answer on training rights confirmed in writing.
  • De-identification as staff were told: names and contact details removed or masked; personal data in tickets and CRM removed, through the de-identification pilot the owner asked for or the owner's own tooling, then our privacy review.
  • Counsel's view on licensing India-based staff's records under India's data protection law before delivery.
  • Exclusivity scope set against the code: once any buyer licenses the code, a whole-company exclusive is off, so archive and code quotes are compared before anything is signed.

Left out

  • HR and payroll records
  • The product's customer data and file storage, and other client and end-user data
  • A chat workspace that includes outside members
  • Third-party prospect data
  • AI chat logs
  • Client contracts and NDA material
  • Payment data, privileged legal material and credentials
  • Code (its own lot; quoted together on request)
  • Names and contact details (removed or masked before delivery)

Lot MKT-ARCHIVE

Exclusive option: about 11 years of a marketing software company's records, 12 tool kinds, about 200 staff

Email, calendar, team chat, documents, work tracking, CRM, support desk, meeting recordings and QA logs from an operating software company of about 200 people, never licensed before according to the owner. Offered with or without the same company's measured code (404 repositories, 266,812 commits).

staff (owner's form)
about200
of records on the longest-running tools (owner's form); years per tool below
11 years
tool kinds on the owner's form; the public knowledge base is not part of the license, and code is lot MKT-CODE
12
countries the records come from
4
earlier licenses of these records (owner's form)
0

What is inside

  • Tool kinds per the owner: email, calendar, team chat, documents, work management, code, CRM, support desk, meeting recordings, knowledge base, QA and test logs, and custom systems.
  • The code host holds more than 10 years of work according to the owner; the measured history starts in 2011 and is offered as lot MKT-CODE.
  • The support desk: a support chat in use for under a year, whose counter shows 143,623 conversations, and which the owner puts at about 5,000 conversations a month now, with a median conversation length of 8 messages; before it, a self-hosted support system, probably from the company's start (almost 10 years), whose ticket list shows about 1.2 million entries (the owner calls them messages), part of them service messages and alerts, and cleaned more than once, in the owner's words. The counter and the ticket list are on screenshots the owner sent on 30 Oct. The knowledge base is public, so it is not part of the license.
  • Languages, in the owner's words of 28 Oct: mail, the older team chat and the work tracker are mainly in two Eastern European languages, with English after them, and the main team chat lists the same two and English; mail also has Portuguese and Spanish, and the chats perhaps a little; the CRM is probably Portuguese first; the support chat runs in five languages, English among them. English comes first in none of the tools the owner ranked; the inventory measures the share per tool.
  • Software industry, no regulated data, an operating company open to an exclusive license of the whole archive.
  • Years per tool, in the owner's words of 28 Oct: about 200 mail accounts, kept since spring to summer 2022 (the owner is not sure the earlier mail, on another provider, is still there); the main team chat since April 2024, and an older chat, used less since then, with several dozen channels over probably 10 years; the work tracker, always in use, with everything about development in it; the current CRM for about two years, another before it, with client data, statuses and notes. Sizes come with the owner's inventory before any contract.
  • From screenshots the owner sent on 30 Oct: the main team chat's statistics page shows 9,541,297 posts in 881 channels and 41 teams, with 223 activated users; the work tracker's current instance runs from June 2017 to October 2026, with issues numbered up to at least 75,800; the company's workspace storage (mail, files and other) has 4.3 TB in use; the CRM runs as two regional accounts: the first one's lists show 408,669 contacts, 35,937 sales email conversations and 12,364 activities still to do, and the second one's counts follow from the owner.

Best for

  • Enterprise agents that work across email, chat, tickets, CRM and documents
  • Customer-support agents trained and evaluated on the company's support desk records (masked; 143,623 conversations on the current chat's counter, and about 1.2 million entries in the ticket list of the system before it)
  • Seed data for workplace RL environments across the licensed tool kinds
  • Meeting understanding from the recorded meetings, subject to a consent review
  • Models of engineering and the business around it, when taken with the code lot

Before delivery

  • A measured inventory per tool (product, years, size, languages), agreed in writing before contract.
  • Records any third party claims rights over are removed or cleared before delivery.
  • Customers' own content (their campaigns, contact lists and their end users' messages) stays out unless customer terms allow it; personal data is masked, and accounts whose contracts forbid use are left out.
  • Signed by the company entity that holds the records, with written assignment from the contractors whose work they are.
  • A privacy review for each of the four countries, and staff notice where the law requires it.
  • Evaluation samples restricted to evaluation, deletion on request; onward transfer and resale terms set in the license; the owner may veto direct competitors.

Left out

  • Customers' own content unless customer terms and opt-in allow it
  • Records under a third party's training-rights claim
  • Personal data (masked) and any regulated data (none per the form)
  • Code, unless the buyer takes the whole-company exclusive (offered as MKT-CODE)

Lot RE-2

About 1.9M outbound emails with 17K labeled replies and 4,100 booked meetings, plus the company's sales scripts and playbooks (owner's count)

One US software company's outbound email record: sends, about 9.7M delivery events, labeled replies and booked meetings, plus campaign and pipeline records and about 6M words of its own SOPs, sales scripts and playbooks. A model learns how real prospects answer cold outreach, with each reply's label as the target. For scale, one open research set holds 1,857 real labeled replies, and a public reply classifier was trained on 5,000 synthetic ones; this record holds about 17,000 by the owner's count.

outbound email sends (owner's count)
About1.9M
labeled replies, from 14K+ repliers (owner's count)
About17K
booked meetings (owner's count)
About4,100
email delivery events, with campaign and pipeline records (owner's count)
About9.7M
internal SOPs, sales scripts and playbooks (owner's count)
About6M words

What is inside

  • Outbound email: about 1.9M sends, more than 14,000 repliers and about 17,000 labeled replies, by the owner's count.
  • The app database: about 9.7M email delivery events plus campaign and pipeline records, by the owner's count.
  • About 4,100 booked meetings, by the owner's count; whether each one links to the email thread that booked it is to be stated.
  • An internal knowledge base of about 6M words: SOPs, sales scripts and playbooks, by the owner's count.
  • Help-desk conversations with the company's customers are also on offer; count and years to be stated.
  • Part of the record is the company's customers' campaign data, offered as far as its customer terms allow and anonymized, after our review of those terms.
  • Prospects' names, addresses and email addresses are masked, and reply text goes through a privacy review before delivery.
  • The same company's consented sales demos are a separate lot (RE-1), so a buyer can pair outreach with live calls.

Best for

  • Classifying and routing real cold-email replies, with the record's own labels
  • Training AI SDR agents to pick the next step after a reply, with booked meetings as the outcome where linked
  • Deliverability and send-pattern models from about 9.7M delivery events
  • Grounding and evaluating sales agents on a real company's SOPs, scripts and playbooks
  • Held-out test sets of real replies for outbound-email tools

Before delivery

  • The owner states years covered, the distinct-message count after templates, the reply label set and who labeled (a person or a model).
  • Which outbound messages a model wrote is stated before delivery, so buyers that exclude model-written text can filter.
  • We and our counsel read the customer-terms clause before any customer campaign data is included.
  • Prospects' and homeowners' names, addresses and email addresses are masked; reply text passes a privacy review; US state privacy rules are checked.
  • Each booked meeting is linked to its thread where the data allows, and the share that links is stated.
  • The company's state registration is confirmed before contract.

Left out

  • Prospects' and homeowners' names, addresses, email addresses and phone numbers (masked)
  • Customer campaign data that the company's customer terms do not cover
  • Replies that fail the privacy review
  • Customers' personal details in help-desk conversations (masked)

Lot REC-BROADCAST

A European broadcast-software company with 20 years of history: company records and its product code

An operating software company for television broadcasters, with 54 staff today and 125 at its peak: email, chat, documents, work tools, support and in-house systems from a 20-year-old company, plus the product code behind three software ranges.

Of company history (owner's form)
20 years
Staff today (125 at the peak)
54
Kinds of tools: email, chat, docs, work, code, support, in-house
7
Lines of product code, over 1,000 commits (owner's form)
Over200,000

What is inside

  • Records in English from one company, with staff told of the use.
  • Three independent product repositories by the owner's email (the form listed 12), self-hosted, with over 100 merged pull requests (owner's form); vendor SDK and codec code inside them is listed or left out.

Best for

  • Enterprise-agent training on a long-running B2B software company
  • Coding agents on broadcast and media software

Before delivery

  • Customer and personal data are masked before delivery.
  • The developers' rights in the code are confirmed before signing.

Lot REC-AGENCY

An Australian marketing-services agency: 12 years of chat, tickets, support and SOPs, plus documents

How a marketing-services agency runs: chat, project tickets, a support desk and a knowledge base kept for 12 years, plus documents and 8 years of in-house systems. 23 staff today, 42 at the peak.

Of chat, tickets, support and wiki
12 years
Staff today (42 at the peak)
23
In English
About75%

What is inside

  • Workspace admin access and chat export are in place.

Best for

  • Agent training on agency and client-service work
  • Support and SOP data for operations agents

Before delivery

  • Deliverables owned by the agency's clients are left out or cleared with them.
  • Personal data is masked before delivery.

Lot REC-JOBS

An Eastern European job and career platform: 10 years of company records

A job and career platform with 25 staff (30 at the peak): email, calendar and documents for 8 years, work tools for 6, CRM for 5 and chat for 4. Personal data of candidates and employers is masked.

Of operation, history kept
10 years
Staff today (30 at the peak)
25
Kinds of tools, from email to CRM
6

What is inside

  • Records from three countries; a quarter or less in English.

Best for

  • Recruiting and marketplace agents
  • Multilingual business records, a quarter or less in English

Before delivery

  • A privacy review comes first; candidates' and employers' personal data is removed or masked.

Lot REC-CONSULT

A US professional-services firm winding down: 8 years of email, calendar, chat, documents and meetings

A 10-person US services firm closing its doors, with its full history kept: email, calendar, chat, documents, work management, meetings, a knowledge base and its data warehouse, all in English.

Email, calendar, chat and documents
8 years
Staff
10

What is inside

  • One entity, US records, all in English.

Best for

  • Agents for professional services
  • A complete small-company record set

Before delivery

  • Regulated and personal data are removed or masked before delivery.

Lot REC-HARDWARE

A UK hardware and firmware company winding down: 16 years of finance, contracts, designs and firmware

A UK engineering firm founded in 2010, 45 to 49 people, about 2 TB in English: finance from 2010 to 2026, contracts and its product IP, from firmware to hardware designs, schematics and approvals, as described by its introducer and not yet verified.

Of records in English
About2 TB
Finance and contracts
2010 to 2026
People
45 to 49

What is inside

  • Firmware, hardware designs, schematics and design files with their approvals.

Best for

  • Hardware and firmware engineering agents
  • A complete engineering-company record set with its IP

Before delivery

  • The owner's legal name, its status and the signer's authority to license are confirmed before samples.
  • Staff personal data is left out; designs pass an export-control screen before delivery.

Lot REC-SOLAR

An American installer of solar panels and home batteries: several terabytes of email, documents, CRM and project histories

An operating installer that also offers security and home services: email, documents, CRM activity, project histories, proposals and training materials, from a team of about fifty employees and contractors.

Of records
Several TB
Employees and contractors in the past year
About50

What is inside

  • Proposals and project histories, with operational and training materials.

Best for

  • Field-service and sales agents
  • Home-energy project data

Before delivery

  • Customer personal data is masked; contractors' records are included only where the company owns them.

Lot REC-RETAIL

A US consumer brand's retail and marketing operations, 2016 to now: more than 42,000 emails and nearly 19,000 files, de-identified

How a consumer brand reaches major retailers and marketplaces: about 10 years of business email, retail, marketing and operating documents and fulfillment records, offered with customer personal information removed.

Emails in four business mailboxes (orders, press, the main brand mailbox and a staff mailbox), 2016 to now, by the owner's counts, before creator, influencer and community mail is taken out
More than42,000
Files in the company's shared file store, by the owner's count
Nearly19,000
Of retail, marketing and operating documents
About10 years

What is inside

  • Fulfillment records for major US retailers and an online marketplace, as far as the owner's retailer and supplier agreements allow; the owner is checking them.
  • Email from 2016 to now, by the owner's counts: 3,000 in the orders mailbox, 10,500 in the press mailbox, 18,650 in the main brand mailbox and more than 10,000 in a staff mailbox. Correspondence with creators, influencers or community members in them stays out.
  • Two smaller mailboxes are checked before they are counted, and the founder's mailbox is not part of this lot until it is filtered to business mail.

Best for

  • Retail and e-commerce operations agents
  • Email agents for order and press correspondence

Before delivery

  • Only the de-identified portion is offered; customer personal information stays out.
  • Anything about people under 18 is left out before samples.

Lot REC-SUPPLY

A small US firearm-accessories brand's supplier records, December 2020 to August 2026: pricing email, about 400 procurement records and SOPs

The procurement side of a small consumer brand (accessories only, no firearms): supplier pricing and negotiation email, about 400 purchase orders, quotes and price lists, and 34 internal SOP and product-FAQ documents of about 350 pages. No customer data.

Purchase orders, quotes and price lists
About400
SOP and product-FAQ documents, about 350 pages
34
Supplier email
2020 to 2026

What is inside

  • Suppliers are not named; a sample folder is ready for view-only access.

Best for

  • Procurement and negotiation agents

Before delivery

  • Before any offer, the buyer confirms it takes records from this category.

Lot REC-STARTUP

A small health-software startup's working records: 3 years of email, chat and work tools, patient records excluded

The internal email, chat and work-tool history of a 10-person software startup in health technology. Patient and health records are not part of the lot.

Email, chat and work tools
3 years
Staff
10

What is inside

  • Records in English; staff told of the use.

Best for

  • Startup operations data for workplace agents

Before delivery

  • No patient or health records are offered.
  • A software vendor's claim to training rights over some records is settled before terms.

Lot REC-STUDIO

A 20-person animation and video studio's working records: up to 2 years of email, chat, documents, CRM, meetings and work tools

The operating history of a small creative studio across 8 kinds of tools, with full history kept, including its internal procedures and workflows. Personal data is masked and direct messages are left out before delivery.

Full-time staff today, 25 at the peak, by the owner's answers
20
Of records per tool, by the owner's answers
Up to2 years
Kinds of tools
8
Records in English, by the owner's answer
About75%

What is inside

  • Two years of documents, work and project tools and CRM; one year of email, calendar, chat, meeting tools and knowledge base, by the owner's answers.
  • Full history kept, by the owner's answer.
  • Internal procedures and workflows are part of the records, by the owner's description.
  • The studio's project files and recorded client calls are listed as separate lots.

Best for

  • Workplace agents for small creative and production teams
  • Client work end to end: CRM, email, meetings and project tools in one company's history

Before delivery

  • Personal data is masked, and direct messages and HR channels are left out, before delivery.
  • Some records fall under regulation, by the owner's answer: those are de-identified or left out.
  • Client data held in the CRM, email and meetings is shown only with client names and contact details removed.
  • Staff are told of the use where the law asks for it, before delivery.

Lot REC-EVENTS

An event business's operating playbooks and aggregate metrics, 2025 to 2026: bookings, workflows, AI use, funnels, campaigns and payments

How an event business runs, from its own systems: host and concierge guides, event-planning procedures and communication templates, with aggregate metrics on bookings, workflow steps, AI API use, web and product funnels, email campaigns and payments. Personal records, correspondence, transcripts and media stay out, as the owner proposes.

Production database tables, 142 of them populated, by the owner's census
193
AI API call records with model, tokens, cost and workflow labels and no prompt text, May to August 2026, by the owner's census
166,360
Document candidates in the company's live files, January 2025 to October 2026, by the owner's search (for review, not all licensable)
225
Web analytics events, January 2025 to October 2026, with about 897,000 sessions (aggregate totals)
About3.5M
Product analytics events across 159 event types (aggregate totals)
About1.5M
Live payment charge attempts, January 2025 to October 2026, offered as timing, failure and refund summaries
1,316

What is inside

  • Operating documents: a September 2026 export of 105 documents and spreadsheets, with host and concierge guides, communication templates, onboarding, event-planning and outing procedures, email templates and consultation scripts among them by file name, and 225 candidates in the company's live files. A content check of three of them found 8,860 words of operating instructions, by the owner's account.
  • Production aggregates, by the owner's census: 16,605 workflow steps, 8,891 booking events, 675 event-service deliveries, 77,579 nurture email sends and 3,781 email-edit observations, 2026.
  • Email marketing: 12 flows, 9 of them live, 28 email campaigns and 22 SMS campaigns, offered as their structure, templates and aggregate outcomes.
  • Payments: of the 1,316 charge attempts, 1,095 succeeded and 221 failed; 77 successful charges were later refunded in part or in full.
  • Sources overlap: one booking, purchase or message can appear in several systems, so no combined total is claimed, and the counts describe retained sources, not final delivery volume (owner's words).

Best for

  • Operations agents for small service businesses: playbooks, templates and follow-up flows
  • Evaluation sets on AI-assisted operations: AI API use, cost and workflow metrics from a live business
  • Booking, funnel and campaign analytics from one business, 2025 to 2026

Before delivery

  • Only nonpersonal operational material and aggregate metrics are offered, as the owner proposes: raw customer or employee records, contact lists, private correspondence, call transcripts, recordings and other media, credentials, and purchased or imported third-party data are left out.
  • Content, ownership and the platforms' terms are reviewed before samples.
  • Samples and final terms remain subject to the owner's written approval.

Lot REC-DEV-STUDIO

A Portuguese-language software studio: 1.1 million logged coding-agent events, with its chat, email, sales calls and code

How a small software studio works with AI coding agents: logged agent events with the failed attempts and the corrections that followed, beside the founder's chat and the team's email, sales calls, notes and documents. The records are in Portuguese; the code is in Portuguese and English.

Logged coding-agent events, mostly from 20 to 28 October 2026 (one agent's log starts 26 September), plus 524,474 complete model calls (prompt and response)
1,102,336
Messages in the founder's own chat account, almost all from 2026 (the oldest from December 2021)
415,018
Sales calls with full transcripts, since 29 December 2025
715

What is inside

  • Also 15,286 emails from team mailboxes, 1,437 team chat messages, 10,756 written notes (5,850 of them authored prose) and 10,844 documents, including 1,359 PDF, 614 CSV and 4,296 JSON files.
  • The code: 720 repositories (671 private) under the founder's own account and the studio's organization, with 338.1 million tokens of source and 1.99 billion of history, duplicates counted once, and 104,640 commits from July 2018 to October 2026. Ownership is confirmed repository by repository at licensing.
  • The agent sessions keep the failed attempts and the corrections that followed, logged during real work rather than curated.

Best for

  • Coding-agent training on real sessions, failures and corrections included
  • Portuguese-language business chat, email and sales-call data
  • Agents for software-studio and client-service work

Before delivery

  • Each model behind the agent sessions is listed with its terms and checked for this use before signing.
  • Repositories that belong to the studio's clients are left out or cleared with them.
  • Personal data is masked before delivery.

Lot REC-UK-GROUP

A UK group's company records: two businesses since 2020, 32 people, about 1.4 TB

A services group with an environmental business and a media and publishing business, with 6 years in each of 10 kinds of tools: about 25 million business emails by the owner's count, recorded calls, team chat, documents, CRM, support and code, about 98% in English.

Of records, by the owner
About1.4 TB
Business emails, in the owner's words
About25M
Of video
About1,200 hours
Staff; about 98% English
32

What is inside

  • Email, calendar, chat, documents, code, CRM, support, meetings, finance and custom tools, 6 years each, on the owner's form.
  • About 230 users or mailboxes, on the owner's form.
  • About 2,000 support tickets and 85 hours of recordings, on the owner's form.
  • About 115,000 payment records stay out until their fields are known.

Best for

  • Business agents trained on a whole company's records
  • Email and support assistants, mostly in English

Before delivery

  • Staff notice and the regulated records are settled before samples; personal data is masked, and payment records stay out until their fields are known.
  • Client material, including video and recorded calls, goes in only with each client's release and the recorded people's agreement.
  • An exclusive license is open on some of the records only, on the owner's form.

Lot REC-CLOSED-COMPANY

A closed Eastern European company's records: email, chat, documents and CRM with full history, 50 staff at the peak

The records of a company that has stopped trading, kept with their full history: email, chat, documents and CRM, about half in English. A staff policy covers this use, or staff can be notified, on the owner's form.

Kept, on the owner's form
Full history
Staff at the peak
50
In English
Abouthalf

Best for

  • Business agents on a company's working records
  • Business records, about half in English

Before delivery

  • A privacy review comes first; personal data is removed or masked, and regulated records are identified and left out.
  • What the company made or sold, and how many years the records cover, are confirmed before samples.
  • The records have been licensed before, non-exclusively, on the owner's form; what that license covers is checked before any exclusive is offered.

8 lots

Audio and video

Recorded calls, interviews and demos

Lot HRT-INTERVIEWS

AI job interviews, audio and video, licensed only where the candidate opts in to AI training: a planned monthly stream from a live hiring product

Real job candidates answer an AI interviewer on camera. The supplier records each speaker on its own channel, with a transcript and the interview outcome, in English (about 90%) and Indian languages. Only sessions where the candidate ticks a separate, optional box naming AI training are licensed, and buyers can add questions written to their own spec.

interviews a month planned, from free interviews the supplier offers its customers (supplier's figure); the opted-in share is measured once the opt-in runs
about50,000
English; the rest in Indian languages (supplier's figures)
about90%
in 21 sample sessions, measured from the files
5.6 hours
per sample session, 16 on average; 720p video at 30 fps
10 to 27 min

What is inside

  • Real job candidates, screened for jobs at employers in India, answer an AI interviewer in a hiring product that is in production today; the supplier plans about 50,000 interviews a month from free interviews offered to its customers, by its own figure.
  • Each session is audio and video with each speaker on a separate channel, plus a transcript and the interview outcome (supplier's description). The candidate is on camera; the interviewer is an AI voice.
  • Consent: a separate, optional checkbox in which the candidate agrees that the audio, video, transcript and outcome may be licensed to AI developers to train and evaluate AI models; the interview and its result do not depend on it, and it can be withdrawn at any time. The supplier says the box is unticked by default and that every file carries its consent record; Rightmark checks both against the files before any sample.
  • Only opted-in sessions are licensed. Interviews recorded before the opt-in are not offered: their consent covers recording only.
  • Questions written to a buyer's spec can be added to the interviews, so a buyer can order spoken answers to its own question set by the month.
  • Measured sample: 21 sessions, 5.6 hours in all, 10 to 27 minutes each (16 on average), H.264 video at 1280x720 and 30 fps with 2-channel AAC audio at 44.1 kHz. Samples are shared only after each file is matched to its consent record.
  • The supplier says it can email Rightmark each withdrawal made after delivery, so the licensee can delete those sessions.

Best for

  • Speech recognition and understanding on English and Indian-language speech from job candidates in India, with transcripts
  • Voice-agent and turn-taking work: real people in a structured conversation with an AI interviewer, on separate channels
  • Evaluation sets for AI interviewing and spoken-answer assessment, with outcomes
  • Audio-visual speech: on-camera answers under an opt-in that names AI training
  • Custom collection: spoken answers to a buyer's own questions, ordered by the month

Before delivery

  • The opt-in is live in production, and every delivered session carries its consent record: time, wording version, tick state.
  • Each sample file is matched to its consent record before any sample is shared.
  • Candidate pay terms seen: the supplier says candidates are not paid for opting in now, and may be later.
  • The terms of the vendor behind the AI interviewer's voice checked; its channel masked or excluded if they require.
  • Rights under the employers' contracts (who owns the recordings) confirmed in writing; the signing company's registry checked.
  • License for recognition, understanding and evaluation only, with a no-cloning covenant and audit rights; never to voice-cloning or avatar buyers.
  • Withdrawals passed on after delivery, with deletion by the licensee.
  • Counsel on India's consent rules for the opt-in, on any US candidates' video, and on licensing interview outcomes for model training.

Left out

  • Past interviews recorded before the opt-in (their consent covers recording only)
  • Any session without a matching AI-training opt-in record, and any withdrawn session
  • Speech generation, voice cloning, avatars and likeness synthesis

Lot RE-1

B2B sales demos with US real-estate agents under an AI-training notice from 26 October 2026 (owner's plan): about 60 to 75 hours a month at today's rate, each call with transcript, speaker labels and outcome (owner's count)

Real sales demos between a software company's closers and real-estate agents across the US. From 26 October 2026, by the owner's plan, each one is recorded under an AI-training notice given in the invite and said at the start of the call. Each call carries a machine transcript with speaker labels and its logged outcome, so a model learns how discovery and a screen-shared product walkthrough play out, and how each call ended. For scale, one public research set of real B2B sales calls holds 50 calls.

recorded sales demos, by the owner's count; 10 to 25 more hours of client calls on request
60 to 75 hours a month
per call: two-party video calls, closer and prospect
30 to 60 minutes
where the real-estate-agent prospects are, by the owner's statement
50 states
AI-training notice in the invite and said at the start of the call (owner's plan)
Every call from 26 October 2026
each call's outcome logged with the recording (owner's plan)
1 outcome per call

What is inside

  • Two-party video calls between the company's closers and real-estate-agent prospects across the US, 30 to 60 minutes each: discovery questions, then a screen-shared walkthrough of the product.
  • Each call comes with the notetaker's machine transcript and speaker labels. The audio is one mixed track per call, with no separate speaker tracks.
  • Consent is collected at the source (owner's plan, from 26 October 2026): the invite names AI training, the notice is said at the start of the call, on the recording, and the closers sign a release naming AI training. Calls where anyone declines stay out.
  • Every hour in the lot is recorded from 26 October 2026 on. Demos recorded before the notice started are never offered.
  • Each call's outcome is logged; the label set comes with the samples.
  • Screen shares are kept free of homeowner details where possible and flagged where not; a flagged part is cut, or that call is delivered as audio only.
  • Samples: a mock demo recorded the same way, then 3 to 5 demos recorded under the notice, with transcripts; the owner aims to send them in the week of 2 November 2026.
  • Licensed by the month as it is recorded: per hour delivered, batch by batch, non-exclusive unless the owner agrees to an exclusive.

Best for

  • Speech recognition and speaker diarization tests on long, two-party video-call audio with speaker-labeled transcripts
  • Training and evaluating AI sales agents and sales-coaching models on real discovery calls, with each call's outcome as the label
  • Sales-conversation understanding: outcome prediction, next-step and objection detection, call summaries and CRM notes
  • Held-out evaluation sets for meeting assistants that summarize customer calls

Before delivery

  • The AI-training notice is in every invite and said at the start of every delivered call, from 26 October 2026; the wording in use is on file.
  • Each closer's signed release naming AI training is on file.
  • Calls where anyone declines, and every demo recorded before the notice, are left out.
  • Each batch is screened against state recording laws; some prospects are in all-party-consent states.
  • The notetaker's and the video platform's terms are checked for any claim on recordings and transcripts.
  • The license limits use to recognition, understanding, turn-taking and evaluation, with a no-cloning covenant and audit rights.
  • Samples reviewed and hours measured from the recorder export before any invoice; screen shares with homeowner details cut or delivered as audio only.
  • The company's state registration is confirmed before contract.

Left out

  • Every demo recorded before 26 October 2026 (no AI-training notice): never offered, samples included
  • Any call where a participant declines
  • Screen-share segments that show homeowners' details (cut, or the call goes as audio only)
  • Voice cloning, voice synthesis and likeness uses
  • The company's outbound phone calls by its setters (not offered now; two-sided calls may follow later by the owner's plan)

Lot AV-CALLS

About 177,600 hours of inbound US phone calls, April 2022 to October 2026, many with both sides recorded

Raw inbound calls from a US lead-generation business across 11 states: health, Medicare and final-expense insurance lines and Spanish-language debt calls, many live-transferred so both sides are on the recording, with a recorded-line notice played. Calls with health details are held out.

Recorded in all
About177,600 hours
US states
11
April 2022 to October 2026
2022 to 2026

What is inside

  • English and Spanish; raw, not transcribed or redacted yet.
  • Most of the hours are health-insurance calls, which stay out for now; the share without health details is being counted.

Best for

  • Speech and voice-agent models for phone sales and service
  • Spanish-language call audio

Before delivery

  • Calls that carry health details stay out until our health-data program is in place.
  • Personal details are redacted before delivery, and the rights in recordings from the former business are confirmed.

Lot AV-CALLS-CLIENT

About 350 hours of recorded client calls from an animation and video studio, about 597 calls, with transcripts and speaker labels

Calls between a creative studio and its clients, never published, with transcripts, call summaries and separate speaker tracks or labels. People on the calls can be identified and have no consent that covers AI training, so voices and names are de-identified, or such consent is put in place, before any sample.

Recorded calls, by the owner's answer
About350 hours
Client calls, by the owner's description
About597
Transcripts and call summaries, by the owner
Included
Separate speaker tracks or labels, by the owner's answer
Included

What is inside

  • Client calls of a studio that makes commercial animation and video, by the owner's description.
  • Never published, by the owner's answer.
  • The calls are in English, by the owner's answer.
  • People on the calls can be identified and have no consent that covers AI training, by the owner's answer.

Best for

  • Speech recognition and voice-agent models for client and project calls
  • Meeting summarization: calls paired with their transcripts and summaries

Before delivery

  • Voices and names are de-identified, or consent that covers AI training is in place, before any sample.
  • Personal details in the transcripts and summaries are redacted before delivery.
  • Who holds the rights in the recordings is confirmed before terms (the owner is not sure yet).

Lot AV-PODCAST

An English-language expert-interview podcast: about 260 episodes, 115 of them with guest releases that name AI training

Long-form, unscripted interviews with scientists, researchers, founders, authors and other experts on neuroscience, psychology, biology, health, human performance and entrepreneurship, in audio and, for most episodes, video. 115 episodes (80 hours 48 minutes) carry a guest release that names AI training, and the newest also have separate speaker tracks and speaker-labeled transcripts.

Episodes
About260
Released for AI training (115 episodes)
80 h 48 min

What is inside

  • Voice cloning, synthesis and likeness uses are excluded by the releases.
  • Two release versions: from episode 150, guests let the host keep rights to the recordings for monetization or future use in all public distribution, including AI training, without expressly mentioning licensing to third parties; from episode 262, the release also names licensing to third parties. A buyer may take only the later episodes, or the earlier ones where that wording covers its use.
  • From episode 262 (4 episodes so far, about 3 hours 29 minutes), each episode has separate speaker audio and video tracks and timestamped, speaker-labeled transcripts, and every new episode uses this release. These episodes can be quoted as their own tier.
  • Licensed as audio alone, or with video, transcripts and episode graphics, each format priced on its own. Later episodes come as separate, separately priced batches.

Best for

  • English speech and conversation models (expert interviews)
  • Long-form interview data
  • Expert explanation and reasoning in natural dialogue

Before delivery

  • Only episodes whose guest release names AI training are licensed.
  • The intro and outro music (licensed CC BY-NC, which rules out commercial use) and the pre-roll ads are removed before any license.

Lot AV-RADIO

About 3,000 hours of Libyan Arabic radio since 2013: talk shows, listener phone-ins, interviews and multi-speaker discussions

An independent Libyan radio station's own broadcast archive, kept as original mono MP3 with a documented date for each episode: live talk shows, listener phone-in conversations, interviews, and cultural and entertainment programs, in natural Libyan Arabic. Delivered as broadcast audio, with transcription and speaker separation left to the licensee.

Hours of original radio recordings (owner's figure)
About3,000
On air since 2013; each episode dated (owner's words)
Since 2013
Original broadcast audio, kept as MP3 files (owner's words)
Mono

What is inside

  • Live talk shows, listener phone-in conversations, interviews, cultural and entertainment programs, and natural multi-speaker discussions.
  • The archive is offered as it is; the licensee handles any processing, transcription and speaker separation.
  • The owner makes any agreement subject to confirming the rights and permissions for third-party voices and music.

Best for

  • Speech recognition and understanding for Libyan and Maghrebi Arabic, a low-resource dialect
  • Natural conversational and call-in speech in Arabic dialect

Before delivery

  • Rights and permissions for callers', guests' and music content confirmed before samples, as the owner proposes.
  • Music, jingles and background music removed or cleared before delivery.
  • Personal data said on air, such as callers' names and phone numbers, masked.
  • The station, its owners and the signing party screened against sanctions lists before any agreement.
  • The owner's documents confirming ownership of the station and authority to license come before samples.

Lot AV-SPOKEN

About 700 hours of audio content in Hindi, Bengali and other Indian languages, and English, with more added daily

A library of audio content in several Indian languages and English, growing every day. The owner would look first at what needs no guest's consent, from the existing library and from the new content added daily.

Audio content across several languages
About700 hours
New content added
Daily

Best for

  • Speech recognition in Indian languages
  • Speech models for lower-resource Indian languages

Before delivery

  • Only recordings that need no guest's consent are considered at first; voice modeling of a host is agreed with that host separately.

Lot AV-SALES-CALLS

US outbound sales conversations with outcome labels, from a voice platform, AI voice-agent calls among them

Conversations from an outbound calling platform for US business and consumer campaigns, each labeled with its outcome, offered de-identified with transcripts as the main format. The volumes are being checked.

On the conversations
Outcome labels
The main format, de-identified
Transcripts

Best for

  • Sales and voice agents
  • Call outcome prediction

Before delivery

  • Recording consent and the de-identification are reviewed before samples; personal data is removed.

For data owners

Own data like this? List it with Rightmark.

Start with an indicative appraisal: no files or system access needed. It is free, and nothing is licensed until you sign.

26 lots

Footage, photos and music

Drone and first-person video, photo catalogs and music libraries

Lot AV-DRONE

About 501 hours of drone footage, up to 4K, from one US drone business

Aerial footage up to 4K from one owner. Some of it has been published before.

Of drone footage
About501 hours
Resolution
Up to4K

What is inside

  • The owner's form says no identifiable people; event and real-estate shoots may include some.

Best for

  • Video and world models that need aerial views
  • Geospatial and mapping research

Before delivery

  • No footage of combat, military sites or strike damage; sensitive sites go to approved buyers only.
  • Flight logs and the pilot's license number have been asked for; neither is in yet.

Lot AV-AERIAL

Drone footage from an operating US drone business, with flight logs to come with the samples

Aerial video shot by an operating US drone business. No volume is shown until the owner confirms the hours.

Operating drone business, by the owner's description
US

What is inside

  • Hours, resolution and which clips carry flight logs are confirmed by the owner before samples.

Best for

  • Video and world models that need aerial camera motion
  • Mapping and aerial scene understanding

Before delivery

  • No footage of combat, military sites or strike damage; sensitive sites such as power plants, ports and bridges go to approved buyers only.
  • Flight logs and the pilot's license number come with the samples.
  • Footage shot for clients is shown only where the client contract allows licensing, or left out.
  • Private property and people below are reviewed before delivery.

Lot AV-EGO

145 hours of first-person HD video of household and workplace tasks, with about 1,150 more hours in review

Egocentric video filmed from the wearer's point of view, hands only, by about 1,300 contributors in Brazil who passed an ID check and accepted terms that name AI training. Raw video with task, setting and device metadata.

Approved first-person video; about 1,150 more hours in final review
145 hours
Contributors in Brazil, all adults with an ID check
About1,300
Task, setting, duration, resolution, frame rate, device and recording date
Labeled

What is inside

  • Household tasks (cooking, dishwashing, cleaning, laundry, tidying, crafts, yard work, home tech use) and small-business work (salon services, landscaping, construction, textiles, commercial kitchen, bakery, retail), recorded mainly in October 2026.
  • Every video is checked by machine and by eye; footage showing children, other people's faces, documents, cards, screens with personal data or license plates is rejected.
  • About 18 of the approved hours, and about 26 hours still in final review, are also offered elsewhere on non-exclusive terms; no exclusive license has been granted.
  • Hours never offered, licensed or shown as a sample anywhere, as of 1 Nov 2026, at the owner's top grade (1080p landscape), approved and in final review: cooking and dishwashing 32 and 479; cleaning and tidying 16 and 249; laundry 17 and 151; home tech use 4 and 89; crafts 5 and 53; yard work, hobbies, pets, repairs and errands 4.3 and 71; workplace tasks 0 and 58 (salon and beauty 22, construction 11, landscaping and farm work 7, commercial kitchen and food 5.5, trades and workshops 4, textiles and leather 3, commercial cleaning and hospitality 2, retail and warehouse 1.6).
  • The contributors can film new footage to a buyer's brief, including in factories and on construction sites with the site's permission: about 200 accepted hours a week, confirmed per brief. Stereo capture and spoken step-by-step narration are assessed per brief.

Best for

  • Robotics and embodied-AI foundation models
  • Action recognition

Before delivery

  • The releases are checked against the footage before delivery.

Lot AV-TRADES

First-person video of skilled trades: auto repair, carpentry, renovation and firefighter drills, with new footage to order

Head-mounted video of hands-on work from 15 people across four trades: two mechanics, five carpenters, three renovation workers inside residential apartments, and a five-person firefighter crew recording controlled drills. New footage can be filmed on a steady schedule.

Trades filmed
4
People filming
15

What is inside

  • Auto repair, carpentry and woodworking, finishing and renovation, and firefighting drills.

Best for

  • Robotics foundation models that learn from skilled manual work
  • Task and tool-use recognition

Before delivery

  • Everyone filmed signs a release that names AI training; residents' homes are filmed only with their consent.

Lot AV-PHOTOS

21,729 original photos in RAW with EXIF, 2022 to 2026, plus about 250 counterfactual image pairs

One photographer's catalog, shot mostly on two camera bodies and kept in RAW with full EXIF; about 20,000 of these were used once in a private AI project, and a separate set of just over 4,250 photos, by the owner's count, was never shared or licensed.

Photos cataloged, RAW with EXIF
21,729
Counterfactual image pairs
About250
Years shot
2022 to 2026

What is inside

  • 15,195 of the photos are from 2026.

Best for

  • Image-model training with camera metadata
  • Counterfactual pairs for image editing models

Before delivery

  • Photos with identifiable people are left out unless released: the owner's removal from the never-shared set was a tool pass and a skim, not a photo-by-photo review, and in the older archive he puts clearly visible faces at 10 to 20%, to be set apart, so the set is checked again before delivery.

Lot AV-MUSIC

A European sample-pack label's library: tens of thousands of original loops, one-shots, vocals and sound effects, with thousands of construction kits

Loops, one-shots, vocals, sound effects and construction kits from a European label, about 80% made in-house and the rest by freelancers and vocalists under agreements, across trap, EDM, house, lo-fi and cinematic styles.

Original audio assets
Tens of thousands
Construction kits
Thousands

What is inside

  • Genres: trap, EDM, house, lo-fi, cinematic, sound effects and vocals.

Best for

  • Music and audio generation models
  • Sound-effect models

Before delivery

  • Producers' written assignments covering AI training, and vocalists' agreements, are checked before signing.

Lot AV-BASKETBALL

Four seasons of national-league basketball streams: about 450 men's and women's league games a season, about 1 hour 45 minutes each

Full-length streams of a national basketball league's two divisions, men's and women's, recorded at each venue by the clubs and carried on the streaming platform the federation runs with its broadcast partner. The federation holds the league's rights. Some games have a clean feed without commentary; others carry commentary, and the streams show the platform logo and the scoreboard.

League games a season, men's and women's
About450
Per game as streamed, stoppages included
About1 h 45 min
Seasons of footage, since 2022/23
4

What is inside

  • Recorded at source by the clubs; a broadcast partner provides the streaming platform.
  • Some feeds are clean, without commentary; others carry commentary. Streams carry the platform logo and the scoreboard.

Best for

  • Video models that learn team sports and fast multi-person motion
  • Sports analytics: event detection and player tracking

Before delivery

  • Players identifiable in shot are released, or their faces are masked, before delivery.
  • Title to each competition's footage, including the clubs' recordings and the platform's terms, is checked before samples.
  • Licensed music is muted or removed.

Lot AV-OPTIN

About 20,000 production-music songs whose composers opted in to AI training in writing, composition and master cleared in one step

A production-music library whose composers each signed a written addendum that lets the library license their works for the development, training and application of AI in music analysis, processing and synthesis, worldwide and sublicensable. Each composer grants both the composition and the master, so every track clears in one step.

Songs covered by earlier non-exclusive AI-training licenses
About20,000
Composition and master granted by the same composer
One-stop
A signed AI-training addendum per composer
Opt-in

What is inside

  • Earlier AI-training licenses were non-exclusive; a buyer that already holds these songs is matched out before terms.
  • Only songs whose composer signed the addendum and whose main agreement is still active are delivered.

Best for

  • Music generation, analysis and processing models
  • Buyers that need per-composer opt-in consent on file

Before delivery

  • The owner approves each buyer before terms. Composers get a notice of each sublicense: the type of use and their share, usually without naming the buyer.

Lot AV-SFX

About 3 million human-recorded sound effects in 83 UCS categories, with metadata written by people

A sound-effects library kept as recorded, with no re-encoding and nothing AI-generated, from ambience, footsteps and foley to machines, animals and vehicles. About a third is the owner's own recordings; the rest comes from long-standing partner recordists who own what they submit.

Sound-effect files, counted per UCS category
2,999,994
UCS categories, with UCS metadata
83
Owned outright; the rest from partner recordists
Abouta third
Hours of audio
About68,000

What other catalogs ask and earn

€20,000a year, 53,000+ sound effects
A library of more than 53,000 sound effects, listed for machine learning on an AI data marketplace, asks from €20,000 a year or €3,000 a month. It is the only public AI-training price for sound effects we found.
€60,000one-time, same listing
The same listing asks from €60,000 for a one-time purchase, three times its yearly price. It does not say whether the license is exclusive.
Over $100,000a year, single catalogs
By June 2025, some catalogs in one opt-in AI licensing program for music and sound effects were earning over $100,000 a year from its deals (company-reported).
$25M to $50Meach deal, images, video and music
A stock-media library's AI training deals with large tech companies, for hundreds of millions of images, videos and music files together, started at this size; most were later expanded. The report gives no price for the audio alone (April 2024).
No public pricehalf a million to over 3 million files
Larger libraries listed for machine learning, from half a million sound effects to a collection of over 3 million audio tracks, samples and sound effects, publish no price and quote on request.

Public asking prices and company-reported figures for other catalogs, checked 27 October 2026. We found no public paid deal for a sound-effects library alone. None of these is an offer or a price for this lot, whose terms are set on request: a price for a library of more than 53,000 files does not scale to one of about 3 million.

What is inside

  • Environment and everyday sound: ambience 120,769 files, footsteps 190,328, objects 117,506 and foley 82,509.
  • Machines 59,270 and mechanical 34,021; animals 69,603, birds 20,499 and animal 1,344; vehicles 227,561.
  • Recorded and described by people; the original files are kept as recorded.
  • Sample rate and bit depth: 56% of files at 48 kHz/24-bit, 41% at 96 kHz/24-bit and 2% above that; 0.3% are below 48 kHz/24-bit.

Best for

  • Sound-effect and audio generation models
  • Noise suppression and source separation, from ambience and background sound

Before delivery

  • For the partner recordists' share, the written grant that lets the owner license their recordings for AI training is shown before terms.
  • Voices, human and crowd sounds (155,578 files) are delivered only with performer releases that name AI training.
  • The owner approves each buyer, and buyers it already licenses effects to directly are excluded.

Lot AV-TRACKS

About half a million instrumental music tracks with metadata, some with separate stems

Instrumental music from an established production-music catalog, every track with metadata; some tracks come with stems and the rest are stereo.

Instrumental tracks, each with metadata
Abouthalf a million

What is inside

  • Stems for some tracks; stereo for the rest.

Best for

  • Music generation models
  • Music tagging and analysis models

Before delivery

  • Composition and master rights, and any session musicians' agreements, are mapped for every track before terms.

Lot AV-SONGS

About half a million vocal music tracks with metadata, about 4,000 with separate instrument stems and vocal tracks

Songs with vocals, nearly all in stereo, every track with metadata. About 4,000 come with true instrument stems and separate vocal tracks, which are rare.

Vocal music tracks, each with metadata
Abouthalf a million
With instrument stems and separate vocal tracks
About4,000

What is inside

  • Nearly all tracks are stereo mixes.

Best for

  • Song and singing-voice generation models
  • Vocal separation, from the multitrack songs

Before delivery

  • Composition and master rights are mapped for every track, and singers' agreements are checked for AI training, before terms.

Lot AV-RENDERS

Audio and MIDI pairs rendered to order from deep-sampled virtual instruments, in any key, tempo, layer or stem

Millions of samples from deep-sampled instruments and short arrangements made by people inside commercial virtual-instrument products, rendered to a buyer's brief with the matching MIDI.

Instrument samples to render from
Millions
Paired, in any key, tempo, layer or stem
Audio and MIDI

What is inside

  • Arrangements were written by people; renders come with their MIDI.

Best for

  • MIDI-conditioned and symbolic-to-audio music models
  • Instrument and timbre models

Before delivery

  • The render brief (instruments, keys, tempos, layers and volume) is agreed before terms.

Lot AV-MOTION

About 250 commercial animation and video projects from one studio, about 300 minutes of finished animation with storyboards, rigs and editable project files

A working studio's projects kept whole: the brief, scripts, storyboards, editable motion-graphics and 3D files, revisions with client feedback, and the final video. 2D and 3D work. The figures are the owner's own; samples come after the owner confirms the lot.

Commercial projects, by the owner's description
About250
Finished animation, by the owner's answer
About300 minutes
Projects with editable After Effects files, by the owner's estimate
About230 to 240
Storyboards and design files, by the owner's description
About200
Both styles, by the owner's answer
2D and 3D

What is inside

  • Each project as the owner describes it: the brief, scripts, storyboards and design files, editable motion-graphics project files, some 3D files, revisions and client feedback, and the final video.
  • Source files kept with the finished work, by the owner's answer: storyboards and animatics, line art, model sheets and keyframes, character rigs and 3D models, and layered scene and project files.
  • By the owner's estimate of 28 October: about 230 to 240 of the projects keep their editable After Effects project files and about 200 have storyboards; scripts and detailed briefs exist for about 150; about 100 to 150 carry the whole chain, from the client's brief or call through script, storyboard, editable files and revisions with feedback to the final video.
  • Two complete projects the owner shared as folder lists on 29 October, names removed, follow one layout: the project brief, the script and visual direction, and briefs for the editor and the designer; the source material (footage, sound, images, icons, fonts, UI and logo files); five or six revision cuts; the final video, about 68 MB and 98 MB; and the editable After Effects project archive, about 150 MB and 1.5 GB.
  • Made mainly in After Effects, with Blender and Cinema 4D for the 3D work, and client-facing material (scripts, briefs, calls and feedback) in English, by the owner's answers. Final videos run about 50 to 400 MB each; the editable archive, with its linked assets, design files, footage, audio and renders, is larger, and its size waits for the owner's storage audit.
  • The studio also holds recorded client calls and its working records; those are listed as separate lots.

Best for

  • Video and animation generation models that learn from source files, not only finished frames
  • Creative-workflow agents: from brief and script through storyboard and revisions to the final cut
  • Storyboard-to-video, rigging and scene-layout research

Before delivery

  • Work made for clients is offered only where the studio holds the right to license it for AI training, or with the client's written consent; the rest is left out.
  • Client names, logos and contact details in briefs, scripts and feedback are removed before delivery.
  • Stock footage, music, fonts and other third-party material inside the project files is listed and left out before delivery.
  • Copies are shared for review; the studio keeps its masters.

Lot AV-SESSIONS

200 vocal songs in full session detail: dry and processed vocals, DI and amped guitars, MIDI, lyrics, chords and effect chains

Songs built from scratch as training data and delivered like a working session: dry and processed vocals, DI and amped guitars, virtual-instrument MIDI with audio renders, lyrics, chord progressions, and the effect chains and automation behind each mix. 200 songs are done; the owner plans about 10,000 within about 18 months.

Songs completed
200
Planned within about 18 months
About10,000

What is inside

  • The owner is looking for partners for the build.

Best for

  • Music generation with control over stems and the mix
  • Mixing and vocal-processing models

Before delivery

  • The singers' agreements name AI training up to voice blending; voice cloning, a 1:1 reproduction of a singer's voice, is not included and is negotiated separately.

Lot AV-FILMS

A European world-sales company's catalog: about 200 films and series, represented for all rights, AI included

About 200 finished titles curated over two decades, in versions of various lengths (one-hour, feature and series), which the company says it represents for all rights, AI included.

Finished titles, in one-hour, feature and series versions
About200
Represented, AI included, in the owner's words
All rights
Hours of film in the catalog, per the owner
About500
Broadcast-quality masters, many in 2K or 4K
HD or better

What is inside

  • Most titles are films the company sells for their producers; about a dozen it produced itself, and their materials beyond the finished film are lot AV-FILM-RAW.
  • Every film has English subtitles and many have English narration; other languages come as separate SRT subtitle files, depending on the film.

Best for

  • Video models trained on professionally made long-form film
  • Film and series understanding and captioning

Before delivery

  • Each title's sales agreement is checked for the AI-rights mandate before signing; titles without it stay out.
  • Hours, formats and resolution are confirmed title by title.

Lot AV-FILM-RAW

About a dozen films a company produced itself, with all their material beyond the finished film

For its own productions the company holds every material outside the finished film, so each of these titles brings more hours than its final cut.

Own productions
About12
Outside the finished film, held by the producer
All material

What is inside

  • The owner also counts roughly 500 hours of rushes, less prepared, tagged and cataloged than the finished films; how much of it belongs to these productions, and the kinds of material, are confirmed title by title. The finished films themselves are in lot AV-FILMS.

Best for

  • Video models: professional material outside the finished film
  • Editing and story models: material beside its final cut

Before delivery

  • Releases from the people filmed are checked before samples; material without them stays out.

Lot AV-POV-HOME

About 350,000 hours of first-person household video from five countries, head-mounted

Ready-made video of household tasks (packing, appliances, organizing, furniture care) filmed from the wearer's point of view on head-mounted iPhones in India, Indonesia, Nigeria, Kenya and the Philippines. 1080p at 30 frames a second, with no action labels or hand-pose annotations.

First-person household video, head-mounted
About350,000 hours
India, Indonesia, Nigeria, Kenya and the Philippines
5 countries
Filmed on iPhone; no action labels
1080p, 30 fps

Best for

  • Robotics and embodied-AI foundation models
  • Action recognition

Before delivery

  • The consent of everyone on camera and the owning company's rights are checked before any sample or license.

Lot AV-POV-FACTORY

About 68,000 hours of first-person factory work video with IMU motion data

Head-mounted first-person video of factory work (assembly, garments, pharma, electronics, metalwork, inspection and packing), filmed on a custom head rig at 1080p and 30 frames a second with IMU motion data. No action labels or hand-pose annotations.

First-person factory work video
About68,000 hours
Motion data recorded with the video
IMU
Custom head rig; no action labels
1080p, 30 fps

Best for

  • Robotics and embodied-AI foundation models
  • Action recognition

Before delivery

  • The consent of everyone on camera and the owning company's rights are checked before any sample or license.

Lot AV-POV-SETS

About 6,800 hours of first-person work and household video in eight smaller sets, several with IMU or stereo capture

Head-mounted first-person video in smaller sets: auto service, repair, cleaning, food prep and two-wheeler production (about 1,300 hours); warehouse, retail and electrical repair on four cameras with IMU (about 1,200); stereo two-lens capture with IMU, household and commercial (about 2,450); construction, renovation, woodworking and handyman work (about 700); household, handyman and hobby tasks with spoken narration and IMU (about 620); electronics manufacturing, bamboo craft and household tailoring, cooking and laundry (about 510 together).

Across eight smaller first-person sets
About6,780 hours
Motion data, two-lens capture or spoken narration in three of the sets
IMU, stereo, narration
Mostly 1080p at 30 fps; the stereo set at 1920x1200
1080p and up

Best for

  • Robotics and embodied-AI foundation models
  • Action recognition

Before delivery

  • The consent of everyone on camera and the owning company's rights are checked before any sample or license.

Lot AV-FIELD-SFX

About 50 hours of location sound effects and ambiences from a working sound editor, with more to order

Effects and ambiences recorded on location by one sound editor, who can record more, quickly, to a brief.

Location recordings, by the owner
About50 hours
More recorded on request
To order

Best for

  • Sound-effect generation models
  • Sound event recognition

Before delivery

  • The recordings are confirmed as the owner's own before samples.

Lot AV-PRINTS

About 5,000 original textile prints and patterns in layered PSD, from 15 years of a design studio's work

A European design studio's library of prints, patterns and graphics for fashion and lifestyle, every file with its original creative layers and some with seamless repeats. Each design is built from the studio's own photography, illustration and hand drawing.

Original designs in the studio's library
About5,000
Layered source files; some with seamless repeats
PSD
Of the studio's work
15 years

What is inside

  • In the studio's words, a design a brand buys is removed from its database at once, so it is never sold to another client.

Best for

  • Image generation for textile and surface design
  • Pattern and repeat generation

Before delivery

  • The studio confirms every design is its own original work, free of third-party material, before samples.
  • Offered as one exclusive deal, with the usage rights transferred to that buyer under its own signed agreement; no non-exclusive copies.

Lot AV-DOCS

Thousands of hours of documentary footage in various formats, from one owner

Documentary video footage one owner holds, in a range of formats. Hours, formats and the rights in each project are to be confirmed.

Documentary footage, held by the owner
Thousands of hours

Best for

  • Video generation models
  • Video understanding on long-form documentary footage

Before delivery

  • The rights in each project and the releases of people on camera are checked before samples; footage owned by others is left out.

Lot AV-VIDEO-ARCHIVE

About 5,700 hours of video in one language, all of it published before, with metadata

Video held by its owner, every hour of it published before and described by metadata, on the owner's form. People appear without signed releases, so video with identifiable people is cleared or left out before any license.

Of video, on the owner's form
About5,700 hours
All of it, with metadata
Published

Best for

  • Video generation models
  • Video and language models on described video

Before delivery

  • People on camera have not signed releases: video with identifiable people is cleared or left out before samples.
  • Resolution, the kind of video and the owner's rights in it are confirmed before samples.

Lot AV-DRONE-NEPAL

About 2 hours of raw 1080p drone video of Nepal, with GPS and height records

15 clips a pilot and team flew over Nepal's cities, night views, heritage sites and landscapes mostly on small DJI drones, kept raw and unedited, with the flights' height and GPS records. The team can fly more to order.

Raw drone video in 15 clips
About2 hours
With GPS and height records
1080p
More flown on request
To order

Best for

  • Video generation with camera motion
  • Aerial scene understanding

Before delivery

  • The raw clips are licensed, not the edited versions with place names written on screen; clips posted on social media are named before samples.

Lot AV-SFX-LIBRARIES

650,000 sounds from more than 300 sound-effects libraries the owner says it owns outright, with detailed metadata embedded in every file

The part of the owner's sound-effects catalog available for AI licensing: 650,000 sounds in more than 300 libraries; the owner says all of it can be delivered at 24-bit, 48 kHz or better.

Sounds available for AI licensing
650,000
Libraries owned outright, by the owner
More than300
Or better, by the owner
24-bit, 48 kHz

What is inside

  • Detailed metadata embedded in every file.
  • The owner says its earlier AI licenses were all non-exclusive, and none of them stops or removes any content from being licensed for AI use.

Best for

  • Sound-effect and audio generation models
  • Sound event recognition and audio tagging

Before delivery

  • The owner says the small amount of third-party content it represents is not included for AI licensing.
  • Recordings with voices are delivered only with releases that cover AI training.

Lot AV-MUSIC-LIBRARY

More than 90,000 royalty-free music tracks, over 1,000 hours, with enriched metadata for each track

Royalty-free music, more than 1 TB of it, with enriched metadata for each track; the owner says all of it can be delivered at 24-bit, 48 kHz or better.

Royalty-free music tracks
More than90,000
Of music, over 1 TB
More than1,000 hours
Or better, by the owner
24-bit, 48 kHz

What is inside

  • Enriched metadata for each track.
  • The owner says its earlier AI licenses were all non-exclusive, and none of them stops or removes any content from being licensed for AI use.

Best for

  • Music generation models
  • Music tagging and analysis, from the track metadata

Before delivery

  • Composition and master rights, and any composers' and performers' agreements, are mapped for every track before terms.

31 lots

Code

Private codebases with their full history

Lot SE-C1

A B2B sales-software company's in-house code: 138 repositories, 4.42 million lines, 118,856 commits over 12 years

The product code of an established B2B sales-software company: a main application with 12 years of history (75,186 commits, 38,849 merged pull requests, 193 contributors) and 137 services, apps, tools and AI components around it. C#, TypeScript, JavaScript and Python with git history, never licensed before by the owner's form, with AI use measured and disclosed per repository.

lines of code in 138 repositories
4,429,008
commits over about 12 years, to October 2026
118,856
merged pull requests
58,938
of history on the main application, 193 contributors
12 years
of commits by an AI-agent identity (2 repositories)
1.9%

What is inside

  • Owner: A B2B sales-software company, about 12 years old, 50 to 120 staff over the years and about 75 today, incorporated in the US with most staff outside the US. By the owner's form, the code has never been licensed, sold or promised elsewhere.
  • Main application: 2,888,550 lines (C# 58.2%, TypeScript 27.9%, JavaScript 11.3%), 75,186 commits and 38,849 merged pull requests over 12 years and 147 active months, 193 contributors, 737,086 lines of tests in 5,394 test files.
  • Around it: 137 more repositories with 1,540,458 lines and 43,670 commits, among them 21 repositories named for AI and LLM work (391,424 lines).
  • Languages over all 4,429,008 lines: C# 47.7%, TypeScript 23.2%, JavaScript 13.6%, Python 10.3%, Swift 1.8%, SQL 1.7%.
  • Engineering practice: CI in 114 of 138 repositories (GitHub Actions in 111), test frameworks in 108, Dockerfiles in 107; 1,244,807 lines of tests in 8,749 test files; 14,840 bug-fix commits, 2,553 of them changed tests in the same commit.
  • Spread over time: 10 repositories started in 2018 or earlier, 94 between 2022 and 2024 and 15 in 2026 (6 of them between August and October); 45 had commits in the 90 days before the check; 55 have 100 or more merged pull requests; contributors per repository up to 193 (median 8).
  • AI use, stated plainly: 2,201 of 118,856 commits (1.9%) come from an AI-agent identity, in 2 repositories, one of them mostly agent-written; 23 repositories carry AI-assistant settings files (180 files); automation and bot commits are being reconciled (between 0.9% and 40% of listed commits by two counts). 50.7% of added lines date from 2023 on, a date measure that cannot show who or what wrote them. 115 repositories (599,482 lines) carry no AI settings and no agent commits, and 9,284,273 of the main application's added lines date from before 2023.
  • Measured by the owner's repository check on full clones: English in 87.3% of 115,498 commit subjects checked; git history complete in 137 repositories, 1 squashed.

Best for

  • Training on in-house enterprise C# and TypeScript with history: 47.7% of the lines are C#.
  • Code review and pull-request models: 58,938 merged pull requests with their commits.
  • Bug-fix and test-writing data: 14,840 bug-fix commits, 2,553 that changed tests in the same commit.
  • Repository-scale context: one application of 2,888,550 lines built over 12 years by 193 contributors, beside 137 services and tools.
  • Software-evolution research: 147 active months of history on one product.

Before delivery

  • Licensed once the owner's IP pack is in: the employment and contractor agreements that assign the code to the company, reviewed by counsel before signing.
  • Every live credential found in the code is rotated before any copy moves, and the owner confirms the rotation in writing by kind and date.
  • The owner confirms in writing which repositories, if any, were ever public; any that were leave the lot or move to the partly-public rate.
  • Personal data in code (2,617 email addresses in 39 repositories) is masked, and data files committed to repositories are removed by path, before delivery.
  • Delivered as git history per repository after these steps; the line count is final after a re-measure without third-party folders and committed data.
  • This lot is a carve-out of the company-records lot: offered on its own only if the records lot is declined or the owner splits the scope in writing.

Left out

  • Repositories that did not clear the provenance review, including public projects, third-party installs and client work.
  • Second copies of a repository's history: each history is priced once.
  • Third-party files inside the repositories (copyright notices in 8 repositories), removed by path.
  • Customer configurations named for outside companies, and personal data in code (masked); every live key is rotated first.

Lot SE-C2

Code rated strong for task building: 13 repositories with 12,889 bug-fix commits, 2,310 of them with their tests, from one B2B software company

Thirteen repositories the owner's repository check rates strong for building coding tasks: 4 pass all five of its checks and 9 pass four. Their 2,310 bug fixes that changed tests in the same commit are the raw material for verified tasks in C#, TypeScript, JavaScript and Python, from code the owner states has never been licensed.

repositories rated strong for building coding tasks
13
bug-fix commits; 2,310 changed tests in the same commit
12,889
lines of tests in 7,720 test files
1,096,669
run CI on GitHub Actions
12 of 13
have no AI-agent commits
11 of 13

What is inside

  • Owner: A B2B sales-software company, about 12 years old, 50 to 120 staff over the years and about 75 today, incorporated in the US with most staff outside the US.
  • The 13: the main application (2,888,550 lines) and 12 more from 11,122 to 342,489 lines, 3,664,431 lines in all: C# 48.7%, TypeScript 24.7%, JavaScript 13.9%, Python 10.1%.
  • Task readiness by the tool's checks: all 13 rated strong; 4 pass all five checks and 9 pass four (the misses: history depth in 6, CI in 1, fixes with tests in 2).
  • Raw material for tasks: 12,889 bug-fix commits, 2,310 of which changed tests in the same commit (1,505 in the main application, 805 in the other 12).
  • History: 91,757 commits and 44,269 merged pull requests; 2 to 193 contributors per repository (median 14); 11 of 13 had commits in the 90 days before the check.
  • Build and test: CI in 12, test frameworks in 13, Dockerfiles in 12; 3 need private packages to build.
  • AI use, stated plainly: 2 of the 13 carry commits by an AI-agent identity (2,201 commits; 1 repository's history is mostly agent commits); 10 carry AI-assistant settings files; 50.7% of added lines date from 2023 on. Without those 2, the subset is 11 repositories, 3,215,495 lines and 1,732 fixes with tests; 3 repositories (143,304 lines) carry no AI settings and no agent commits.
  • English in 87.0% of 88,865 commit subjects checked (lowest repository 74.4%).

Best for

  • SWE-style task sets in C#, TypeScript, JavaScript and Python; SWE-bench draws its 2,294 problems from 12 Python repositories.
  • RL environments for bug-fixing agents, built from fixes that changed tests in the same commit.
  • Data for writing tests and repairing CI: 1,096,669 lines of tests beside the code they test.
  • Evaluation sets from code that, by the owner's form, has not been licensed before.

Before delivery

  • Licensed once the owner's IP pack is in: the employment and contractor agreements that assign the code to the company, reviewed by counsel before signing.
  • Every live credential found in the code is rotated before any copy moves, and the owner confirms the rotation in writing by kind and date.
  • The owner confirms in writing which repositories, if any, were ever public; any that were leave the lot or move to the partly-public rate.
  • Personal data in code (2,617 email addresses in 39 repositories) is masked, and data files committed to repositories are removed by path, before delivery.
  • Each task is priced once it passes the clean-container check; the owner's CI logs and a task-builder pass come first.

Left out

  • Repositories that did not clear the provenance review, including public projects, third-party installs and client work.
  • Second copies of a repository's history: each history is priced once.
  • Third-party files inside the repositories (copyright notices in 8 repositories), removed by path.
  • Customer configurations named for outside companies, and personal data in code (masked); every live key is rotated first.
  • The other 136 repositories of the code lot, offered as the full set (SE-C1).

Lot SE-C3

One B2B software company's code with its engineering record: 138 repositories, issue tracker, engineering chat and support-desk bug reports

The full code lot offered with the records its engineers worked in: the issue tracker (Jira), the engineering channels of a team chat that runs from 2015 to October 2026, and bug reports from the support desk. Where commits and pull requests cite issue keys and issues link the support conversation, a buyer can follow a customer's report to an issue, a fix and a merged pull request; the share that links is measured before contract.

lines of code in 138 repositories, about 12 years to October 2026
4,429,008
merged pull requests; the share citing an issue key is measured before contract
58,938
bug-fix commits
14,840
team chat history; engineering channels selected for this lot
2015 to October 2026
of record in one lot: code, issue tracker, team chat, support desk
4 kinds

What is inside

  • Owner: A B2B sales-software company, about 12 years old, 50 to 120 staff over the years and about 75 today, incorporated in the US with most staff outside the US. By the owner's form, never licensed before.
  • Code: the 138 repositories of the full code lot (SE-C1), with 118,856 commits and 58,938 merged pull requests; the main application alone has 75,186 commits over 12 years.
  • Issue tracker: Jira, the company's tracker (to be confirmed).
  • Team chat: history from May 2015 to 1 October 2026, about 2.18 million messages in 987 channels by the owner's count, not yet measured; the engineering channels are selected for this lot (to be confirmed).
  • Support desk: the conversations that report bugs (to be confirmed).
  • Linkage: commits and pull requests join to issues by the issue keys they cite, and bug reports join where an issue links the conversation (to be confirmed).
  • AI use in the code, as in SE-C1: 1.9% of commits come from an AI-agent identity and 50.7% of added lines date from 2023 on; 115 repositories carry no AI settings and no agent commits.

Best for

  • Agents that work across engineering tools: from a reported bug to an issue, a fix and a merged pull request.
  • Bug triage and root-cause models trained on reports joined to the fixes that closed them.
  • Engineering-workflow environments that mirror a working software company's tracker, chat and repositories.
  • Process data on how one product team planned, discussed and shipped over more than ten years.

Before delivery

  • Licensed once the owner's IP pack is in: the employment and contractor agreements that assign the code to the company, reviewed by counsel before signing.
  • Every live credential found in the code is rotated before any copy moves, and the owner confirms the rotation in writing by kind and date.
  • The owner confirms in writing which repositories, if any, were ever public; any that were leave the lot or move to the partly-public rate.
  • Personal data in code (2,617 email addresses in 39 repositories) is masked, and data files committed to repositories are removed by path, before delivery.
  • Records are de-identified to the buyer's standard and the law's before delivery; customer and prospect data inside chat and support conversations is removed.
  • The owner notifies staff before team chat is shared.
  • This lot is a carve-out of the company-records lot: offered on its own only if the records lot is declined or the owner splits the scope in writing.

Left out

  • Repositories that did not clear the provenance review, including public projects, third-party installs and client work.
  • Second copies of a repository's history: each history is priced once.
  • Third-party files inside the repositories (copyright notices in 8 repositories), removed by path.
  • Customer configurations named for outside companies, and personal data in code (masked); every live key is rotated first.
  • Private channels and direct messages in team chat, and channels outside engineering.
  • Support conversations other than bug reports, and data the company's customers keep in its product.

Lot 02-B

Task-grade code: 14 repositories with CI, test suites and full git history

JavaScript and TypeScript code from one B2B software company, with CI configured in all 14 repositories, more than 6,200 test files and 233 fixes that changed a test in the same commit. Each repository comes with its tests, its CI configuration and its full history: more than 44,000 commits and more than 31,000 merged pull requests in all.

Repositories, CI configured in all 14
14
Original lines with full git history
3.4M+
test files
6,200+
Merged pull requests, counted from merge commits
31,000+
Fix commits that changed a test in the same commit
233

What is inside

  • Full git history in every repository: more than 44,000 commits, first commits 2019 to 2026, none squashed.
  • The largest has more than 2 million lines, more than 17,000 commits, more than 10,000 merged pull requests and 155 people in its history, first commit 2020. 11 of the 14 have more than 500 merged pull requests, the fewest has 54, and 13 of the 14 have 10 or more people in their history.
  • CI: GitHub Actions in all 14, Azure Pipelines in 8, AWS CodeBuild in 5. Frameworks found for the tests: Mocha in 5, Vitest in 4, Jest in 3, Playwright in 1.
  • More than 1.9 million of the lines are in test files, about two thirds of them in the largest repository.
  • The owner reports a test run for each of the 14 with 0 failing and more than 96,000 passing tests in all; three were reported as CI pull-request runs. Rightmark has not seen the run logs.
  • More than 5,600 commits are marked as fixes; 233 of them changed a test file and code in the same commit.
  • Languages by lines: JavaScript 77.7%, TypeScript 22.3%.
  • Disclosed: 72% of commits date from 2023 on, 7 of the 14 commit AI coding-assistant configuration files, and one was built for a client of the owner.

Best for

  • SWE-bench-style environments: a repository at a commit, a failing test and a grader that runs the suite
  • Models that generate and repair tests, on real suites in Mocha, Vitest, Jest and Playwright
  • CI-repair and build agents across GitHub Actions, Azure Pipelines and AWS CodeBuild
  • Agent evaluation on a codebase of more than 2 million lines with 155 people in its history
  • Repository-level code models with full commit history

Before delivery

  • The owner reports that every credential found in the code or its history has been rotated. Credential strings still in the code or its history are removed or shown to be dead before delivery.
  • Any repository found to have been public is left out.
  • The owner's rights are reviewed lot by lot.
  • Once you accept the NDA in Rightmark's buyer portal, Rightmark shares repository-level metrics (names withheld) and the detailed source list. Code samples come only with the owner's agreement.
  • The owner's name, product names and repository names are shared only when the owner agrees to an introduction.
  • One of the 14 was built for a client of the owner. The owner says it may license it; Rightmark has not seen the client contract.
  • 5 of the 14 use private package registries; those packages ship with the code so the repositories can build in a clean container.
  • An exclusive license needs the owner's written OK and takes these 14 out of lot A's non-exclusive sales for its term.

Left out

  • Repositories with an open-source license file (26 in all) and any repository found to have been public
  • Vendored code (more than 110,000 lines in these 14), not counted
  • Customer data in the product's databases and file storage: no code lot carries it

Lot 02-A

A software company's code estate: 266 repositories, 2016 to 2026, with full git history

One B2B software company's git repositories: more than 5 million original lines, more than 88,000 commits and nearly 60,000 merged pull requests from 2016 to 2026, with every repository's history intact. 61% of the commits predate 2023, and 28 repositories commit AI coding-assistant files, so the same company's code spans the years before and after AI assistants.

Repositories
266
Original lines
5.0M+
Commits
88,000+
Merged pull requests, counted from merge commits
Nearly60,000
First commits
2016 to 2026

What is inside

  • Full git history in every repository; none squashed.
  • 61% of commits predate 2023, and 212 repositories have every commit before 2023.
  • Languages by lines: JavaScript 76.6%, TypeScript 20.7%, Python 2.2%.
  • The largest has more than 2 million lines, more than 17,000 commits, more than 10,000 merged pull requests and 155 people; 16 repositories are over 50,000 lines and 138 are under 1,000.
  • CI is configured in 92 repositories (GitHub Actions in 50, AWS CodeBuild in 36, Azure Pipelines in 27, CircleCI in 8), with more than 7,300 test files; Docker files in 117.
  • More than 11,000 commits are marked as fixes; 291 changed a test file and code in the same commit.
  • Disclosed: AI coding-assistant configuration files appear in 28 repositories, and 2 were built for clients of the owner.

Best for

  • Continued pre-training on one company's code from 2016 to 2026, with full history
  • Repository-level and cross-repository understanding across one company's code
  • Commit-message and code-change models over more than 88,000 commits
  • Studying how one company's code changed after AI coding assistants arrived
  • Mining SWE tasks: 291 fix commits that changed a test, more than 11,000 marked as fixes

Before delivery

  • The owner reports that every credential found in the code or its history has been rotated. Credential strings still in the code or its history are removed or shown to be dead before delivery.
  • Any repository found to have been public is left out.
  • The owner's rights are reviewed lot by lot.
  • Once you accept the NDA in Rightmark's buyer portal, Rightmark shares repository-level metrics (names withheld) and the detailed source list. Code samples come only with the owner's agreement.
  • The owner's name, product names and repository names are shared only when the owner agrees to an introduction.
  • Two repositories, one in lot B and one in lot D, were built for clients of the owner. The owner says it may license them; Rightmark has not seen the client contracts.
  • Two repositories with copyleft license headers stay out until the rights review clears them.
  • An exclusive license needs the owner's written OK and cannot overlap a whole-company exclusive that includes the code.

Left out

  • 26 repositories with an open-source license file, and any repository found to have been public
  • 2 repositories with copyleft license headers (more than 660,000 lines), until the rights review
  • Vendored code (more than 2.4 million lines), not counted
  • Customer data in the product's databases and file storage

Lot 02-C

Code from before 2023: 19 repositories with every commit made between 2016 and 2022

Every commit in these 19 repositories predates 2023, and none commits AI coding-assistant files. More than 34,000 commits and more than 23,000 merged pull requests from one B2B software company, with up to 41 people in one repository's history.

Repositories, each over 10,000 lines and 100 commits
19
Original lines
890,000+
Commits
34,000+
Merged pull requests, counted from merge commits
23,000+
First commits
2016 to 2022

What is inside

  • Every commit in every one of these repositories predates 2023. Each has over 10,000 lines and at least 100 commits.
  • All JavaScript, by each repository's main language.
  • 15 of the 19 have 10 or more people in their history, and 9 have more than 1,000 merged pull requests.
  • The largest has more than 100,000 lines, more than 3,500 commits and 28 people in its history.
  • CI is configured in 18 (AWS CodeBuild in 16, GitHub Actions in 5), with 407 test files; Docker files in 12.
  • More than 4,800 commits are marked as fixes; 26 changed a test file and code in the same commit.
  • 3 of the 19 are also in lot B. Another 193 repositories in lot A also have every commit before 2023; they are smaller or have fewer than 100 commits.

Best for

  • Pre-training and evaluation sets of code written before 2023, with no AI coding-assistant configuration files
  • A baseline for studies of AI-assisted code, against the same company's 43 repositories with every commit from 2023 on (lot A)
  • JavaScript repositories with pull-request history for code-change models

Before delivery

  • The owner reports that every credential found in the code or its history has been rotated. Credential strings still in the code or its history are removed or shown to be dead before delivery.
  • Any repository found to have been public is left out.
  • The owner's rights are reviewed lot by lot.
  • Once you accept the NDA in Rightmark's buyer portal, Rightmark shares repository-level metrics (names withheld) and the detailed source list. Code samples come only with the owner's agreement.
  • The owner's name, product names and repository names are shared only when the owner agrees to an introduction.

Left out

  • Repositories with any commit from 2023 on (they are in lots A, B and D)
  • Vendored code (more than 12,000 lines in these 19), not counted

Lot 02-D

16 large codebases, each over 50,000 lines: nearly 4 million lines with CI in all 16

16 codebases of one B2B software company, each over 50,000 lines and with full history: more than 54,000 commits, more than 40,000 merged pull requests, CI configured in every one and more than 6,000 test files. Sized for agents that must navigate and change large, multi-file code.

Codebases, each over 50,000 lines
16
Original lines
3.9M+
Commits
54,000+
Merged pull requests, counted from merge commits
40,000+
First commits
2017 to 2026

What is inside

  • Each is over 50,000 lines; the largest has more than 2 million lines.
  • 7 have every commit before 2023, and 5 were started in 2023 or later.
  • CI configured in all 16, with more than 6,000 test files; Docker files in 9.
  • Languages by lines: JavaScript 78.9%, TypeScript 21.1%.
  • More than 6,900 commits are marked as fixes; 210 changed a test file and code in the same commit.
  • 7 are also in lot B and 7 in lot C.
  • Disclosed: AI coding-assistant configuration files appear in 7, and one was built for a client of the owner.

Best for

  • Long-context coding agents that navigate and edit large repositories
  • Codebase programs that require 50,000+ lines and a container build
  • Multi-file refactoring and migration tasks, with existing tests to check them

Before delivery

  • The owner reports that every credential found in the code or its history has been rotated. Credential strings still in the code or its history are removed or shown to be dead before delivery.
  • Any repository found to have been public is left out.
  • The owner's rights are reviewed lot by lot.
  • Once you accept the NDA in Rightmark's buyer portal, Rightmark shares repository-level metrics (names withheld) and the detailed source list. Code samples come only with the owner's agreement.
  • The owner's name, product names and repository names are shared only when the owner agrees to an introduction.
  • One of the 16 was built for a client of the owner. The owner says it may license it; Rightmark has not seen the client contract.
  • 4 of the 16 use private package registries; those packages ship with the code.

Left out

  • Vendored code (more than 120,000 lines in these 16), not counted
  • Customer data in the product's databases and file storage

Lot 02-E

Code with its context: 30 repositories plus the same company's support desk, engineering chat, work tracking and error monitoring

One B2B software company's code from 2016 to 2026 together with the systems around it: more than 72,000 support tickets, the engineering channels of its team chat, its work-tracking records and recent error monitoring. Agents built for software teams can learn from the code and from the tickets, discussions and work items around it, all from the same company.

Repositories (lots B and C)
30
Original lines
4.2M+
Support tickets
about70,000
Merged pull requests, counted from merge commits
51,000+
Systems beside the code: support desk, engineering chat, work tracking, error monitoring
4

What is inside

  • Code: the 30 repositories of lots B and C, more than 73,000 commits, more than 9,900 commits marked as fixes and 247 that changed a test, first commits 2016 to 2026.
  • Disclosed: 7 of the 30 repositories commit AI coding-assistant configuration files.
  • Support desk: about 70,000 tickets, by the owner's inventory.
  • Team chat: the engineering channels of a workspace of about 60 GB, with about 4.09 million messages in the last 13 months alone; chat history from 2015.
  • Work tracking: a few GB of records; 8 of the 14 task-grade repositories carry pipeline configuration for the same platform.
  • Error monitoring: recent errors only, more than 10,000 in the last 30 days across 12 projects; the owner expects about the last 30 to 90 days are kept. Event payloads, user context, IP addresses, emails and request bodies stay out.
  • Links between tickets, work items and commits have not been counted yet.
  • Delivery only after a rights and privacy review that covers personal data in the tickets, chat and work-tracking records.

Best for

  • Agents that work across an engineering team's tools: ticket, chat, work item, pull request, CI and errors
  • SWE tasks with real problem statements, once a fix is traced to its ticket or work item
  • Support-to-engineering triage and escalation models
  • Error triage with the code and history behind it
  • Evaluating long-horizon engineering work inside one real company's systems

Before delivery

  • The owner reports that every credential found in the code or its history has been rotated. Credential strings still in the code or its history are removed or shown to be dead before delivery.
  • Any repository found to have been public is left out.
  • The owner's rights are reviewed lot by lot.
  • Once you accept the NDA in Rightmark's buyer portal, Rightmark shares repository-level metrics (names withheld) and the detailed source list. Code samples come only with the owner's agreement.
  • The owner's name, product names and repository names are shared only when the owner agrees to an introduction.
  • A rights and privacy review before delivery: personal data in the tickets, chat and work-tracking records is removed before anything is delivered.
  • The engineering channel list is agreed with the owner before delivery; other channels stay in lot F.
  • Lot E's support desk, chat, work-tracking and error data are also part of lot F. If lot F is licensed exclusively, they are carved out before it signs, or lot E waits for that term to end.

Left out

  • HR and payroll, the product's customer data and file storage, a chat workspace that includes outside members, third-party prospect data and AI chat logs
  • Error events' payloads, user context, IP addresses, emails and request bodies
  • Chat channels outside engineering (they are in lot F)

Lot MKT-CODE

A marketing software platform's code and change history since 2011: 404 repositories, 266,812 commits

Its largest repository holds 92,019 commits and 21,559 merged pull or merge requests from 126 contributors over nearly 15 years, and 91% of its added lines date from before 2023. With the 403 repositories around it, the lot carries 266,812 commits and 53,486 fix commits: the record of how a working product was built, reviewed and repaired.

commit history of the largest repository: 14.9 years, with commits in 180 different months
2011 to 2026
original lines of code in 404 repositories (PHP 58%, JavaScript 17%, TypeScript 17%)
7,840,945
commits, with an estimated 75,208 merged pull or merge requests counted from merge commit messages
266,812
fix commits (the subject opens with fix, bug, patch, revert or a similar word); 1,404 changed tests along with the code
53,486
estimated tokens of change history, plus about 104 million tokens of source code
1.89B

What is inside

  • The largest repository: 1,141,718 original lines (54% JavaScript, 45% PHP), 92,019 commits, 21,559 merged pull or merge requests, 126 human contributors, 17,387 fix commits, CI configuration and a PHPUnit suite (217 test files).
  • 101 repositories of 10,000 lines or more hold 7,219,329 lines (92% of the code); 83 of them also have 3 or more contributors, and 25 have 50,000 lines or more.
  • Tests in 179 repositories: 8,586 test files and 923,593 lines of tests; 61 repositories carry CI configuration.
  • 163 repositories (2,185,898 lines) have 10% or less of their added lines from 2023 on, and 33 have 7 or more years of history.
  • 146 repositories had commits in the 90 days before measurement; each repository was measured with its own report on 21 October 2026.
  • Git history is kept in full in 402 of the 404 repositories; the 2 with squashed history are flagged. The owner's form records no earlier license of these records.
  • Set aside before pricing: 15 repositories with open-source license files, 5 that look like imported third-party code, 1 found public and 9 duplicate uploads. The 404 are the cleared candidates among 425 distinct repositories measured.

Best for

  • Pretraining and mid-training on licensed product code with its full change history
  • Code-review and commit-message models from 75,208 merged pull or merge requests
  • Bug localization and repair from 53,486 fix commits in PHP, JavaScript and TypeScript
  • Long-horizon repository understanding: one 1,141,718-line repository followed across nearly 15 years
  • Held-out evaluation sets from product code kept out of public crawls (after the never-public check)

Before delivery

  • Licensed by the company entity that holds the code, with written IP assignments from staff and contractors covering the whole history back to 2011.
  • Any repository that is or was public, or was published as a package, is removed before delivery after a check against public code hosts and package registries.
  • Possible secrets found in the current code are rotated by the owner before delivery; keys are never tested.
  • Files carrying copyleft license headers are removed file by file; email addresses in code are masked.
  • A per-repository disclosure ships with the lot: the share of lines added from 2023 on, and where AI-assistant settings files are committed.
  • Non-exclusive per sale by default; an exclusive license covers the whole cleared set for a stated term.
  • Evaluation samples are restricted to evaluation: no training on them, no retention past the agreed window, deletion on request. Onward transfer and resale terms are set in the license, and the owner may veto direct competitors.

Left out

  • 15 repositories with open-source license files (8 MIT, 5 GPL, 2 Apache)
  • 5 repositories that look like imported third-party code (2,693,131 lines, 11 to 428 commits) until the owner confirms whose code they are
  • 9 duplicate uploads
  • Any repository found to be or have been public
  • Files with copyleft headers, secrets and personal data

Lot RE-3

About 480 AI coding-agent sessions from a working product team, with repositories holding about 3,800 commits and 1,700 pull requests (owner's count)

Sessions in which a software company's team used AI coding agents on its own product, offered with the company's repositories. For contrast, a 258,000-trajectory open release from February 2026 was generated by one model on public-repository tasks; these sessions come from the team's own product work, by the owner's account. Priced on review once the tools, years and session contents are stated.

AI coding-agent sessions (owner's count)
About480
commits (owner's count)
About3,800
pull requests (owner's count)
About1,700

What is inside

  • About 480 AI coding-agent sessions from the company's engineering work, by the owner's count; which agent tools, how far back, and whether each session holds its diffs and test runs are to be stated.
  • Repositories with about 3,800 commits and 1,700 pull requests, by the owner's count; each repository goes through Repo Check on the owner's machine before pricing.
  • Code written by AI agents is labeled as such before any sample, so buyers that exclude AI-written code can filter it.
  • The same company's outbound email record (RE-2) and consented sales demos (RE-1) are separate lots.
  • The owner states all of the engineering data is the company's own.

Best for

  • Coding-agent trajectory training and analysis on real product work, where sessions carry diffs and test runs
  • Repository-level code tasks built from real commits and pull requests, after Repo Check
  • Evaluation of coding agents against how a working team actually used them

Before delivery

  • Repo Check run on every repository, with secrets at HEAD and in history rotated or removed.
  • Ownership and IP assignments from everyone who wrote the code are in hand; public or third-party code is left out.
  • The agent tools' terms are checked for any claim on session logs.
  • Sessions and repositories are labeled for AI-written code before any sample.
  • The company's state registration is confirmed before contract.

Left out

  • Secrets, credentials and customer data found in code or sessions
  • Public, open-source or third-party code
  • Any repository the company does not own outright

Lot VOICE-CODE

Code from an AI voice-agent product, 2025 to 2026: 8 private repositories with full history

One company's own product code for AI voice and chat assistants, written from 2025 on: the agent engine, its platform services, session management, an admin portal and the customer apps. More than 300,000 original lines, 2,447 commits and 2,072 merged pull requests, with 8 to 17 people in each repository's history.

Private repositories, each passing all four repository gates
8
Original lines
300,000+
Commits
2,447
Merged pull requests
2,072
First and last commits
2025 to 2026
People in each repository's history
8 to 17

What is inside

  • The agent engine, in Python: more than 44,000 lines, 539 commits and 284 merged pull requests, with 16 people in its history.
  • A platform-services monorepo, in TypeScript: more than 58,000 lines, 425 commits, 172 merged pull requests and 17 people.
  • Six more TypeScript repositories: a voice and web applications monorepo, an assistant session-management backend, an embedded assistant front end, an admin portal for assistant configuration, a chatbot data-collection service and a generative-analytics customer portal.
  • Test files in 6 of the 8. Some tests were written in September and October 2026 (72 in the agent engine, 25 in the assistant front end), and their dates come with the lot.
  • Disclosed: 7 of the 8 repositories commit AI coding-assistant configuration files.
  • Measured by Repo Check on 19 October 2026, repository by repository.

Best for

  • Coding-agent training and evaluation on a current AI-assistant product: an agent engine, its services and its apps
  • Graded training tasks from the agent engine and the platform services, once their tests run offline
  • Code-change models: more than 2,000 merged pull requests from 2025 and 2026

Before delivery

  • Each contributor's signed IP assignment is reviewed before signing.
  • One repository began as a copy of public projects; that code was removed and its history rebuilt in October 2026, and it comes with that disclosure.
  • Credentials found in the histories were rotated, the owner reports (20 October 2026); the old values remain in the history.
  • A customer's name, which the owner says is not confidential, stays in older commits of two repositories.
  • Task prices apply once the tests of the agent engine and the platform services run green offline.
  • Once you accept the NDA in Rightmark's buyer portal, Rightmark shares repository-level metrics (names withheld) and the detailed source list. Code samples come only with the owner's agreement.

Left out

  • The company's other repositories, which the owner kept out of this lot.

Lot CODE-PROD5

Five private codebases with git history from 2018: 1.2 million original lines in JavaScript, PHP and C#

Five repositories with their full git history, started between 2018 and 2020: JavaScript, PHP, C#, TypeScript and Swift, with 10 to 16 people in each repository's history. Every repository passes Repo Check's four gates for contributors, commits, merged pull requests and original code.

Private repositories with full git history, all passing Repo Check's four gates
5
Original lines (generated and vendored code not counted)
1.2M
Commits
More than4,100
Merged pull requests, estimated from the commit history
More than660
Years of active work (a few later commits, the last in 2026)
2018 to 2022
People in each repository's history
10 to 16

What is inside

  • By language: JavaScript about 430,000, PHP about 353,000, C# about 347,000, TypeScript about 53,000, Swift about 38,000 lines.
  • Two of the repositories grew from the same first commits, so their shared early history is counted once in the commit figure.
  • Tests in 4 of 5 (44 test files; .NET test project, Jest, PHPUnit), and CI on CodeBuild, GitLab CI, Travis CI.
  • At least 640 bug-fix commits, at least 3 of them changing a test (the shared early history counted once).
  • Measured by Repo Check 2.2.0 on 2026-10-19, repository by repository.

Best for

  • Coding-agent training and evaluation on multi-language code in JavaScript, PHP, C# and Swift
  • Code-change models: more than 4,100 commits and at least 640 bug fixes
  • History from 2018: 10 to 27 months of active work in each repository

Before delivery

  • Credentials that Repo Check found in the code are rotated and removed before delivery.
  • Repo Check flagged 8 files with GPL-family license headers and 4 package manifests that could be published to public registries: the owner confirms which parts, if any, are published, and those parts and the GPL-licensed files are left out before samples.
  • Personal data in the code, such as email addresses, is removed before delivery.
  • The owner confirms who holds the rights in each repository before samples and terms.
  • Each contributor's IP assignment is reviewed before signing.

Lot CODE-JS12

Twelve private repositories from one software company, with history from 2017: about 297,000 original lines and 2,400 merged pull requests

A software company's code with its full git history: JavaScript, Java and TypeScript, 8,265 commits and 2,720 bug-fix commits, with 5 to 12 people in each repository's history. All 12 pass Repo Check's four gates, and no credentials were found in any of them.

Private repositories with full git history, all passing Repo Check's four gates
12
Original lines (generated and vendored code not counted)
About297,000
Commits
8,265
Merged pull requests, mostly estimated from the commit history
About2,411
Years of active work (a few later commits, the last in 2026)
2017 to 2025
People in each repository's history
5 to 12

What is inside

  • By language: JavaScript about 244,000, Java about 37,000, TypeScript about 14,000, SQL about 2,000 lines.
  • Tests in 5 of 12 (39 test files; Cypress, JUnit, Jest), and CI on GitLab CI, Cloud Build.
  • 2,720 bug-fix commits.
  • One of the twelve was started in 2023; in the other eleven, under 1% of the added lines date from 2023 or later.
  • Measured by Repo Check 2.2.0 on 2026-10-21, repository by repository.

Best for

  • Code-change and review models: about 2,400 merged pull requests over 12 repositories
  • Coding-agent training on JavaScript and Java code with its history
  • Repair data: 2,720 bug-fix commits

Before delivery

  • Repo Check flagged 5 package manifests that could be published to public registries: the owner confirms which parts, if any, are published, and those parts are left out before samples.
  • The company confirms its rights in the code, including code from before it was incorporated, before samples and terms.
  • Any repository built for a client is shown with that client's written consent, or left out.

Lot CODE-CPP

A game studio's codebase, 2017 to 2026: one monorepo with 1 million original lines, 24,049 commits and 154 contributors

Nine years of a working game studio's code in one repository, in C++, Python, C and Lua, with about 3,386 merged pull requests by the studio's count, 4,000 bug-fix commits and more than 100,000 lines of tests. Too large for the standing rate, so it is priced on review.

Private monorepo with full git history, passing Repo Check's four gates
1
Original lines (generated code not counted; the owner says the repository holds no engine or third-party code)
1M
Commits
24,049
Merged pull requests, by the owner's count
About3,386
First and last commits
2017 to 2026
People in the repository's history
154

What is inside

  • By language: C++ about 433,000, Python about 214,000, C about 171,000, Lua about 132,000, TypeScript about 32,000 lines.
  • Covers game code, backend code, deployment, CI/CD pipelines, QA automation, in-house tools for production teams and the website, as the studio describes it.
  • 651 test files (106,955 test lines; Python tests, pytest), and CI on Jenkins.
  • 4,000 bug-fix commits, 187 of them changing a test.
  • About 19% of the added lines date from 2023 or later; the rest go back to 2017.
  • Measured by Repo Check 2.4.1 on 2026-10-26, repository by repository.

Best for

  • Code-change and repair models on a large C++ and Python codebase
  • Graded training tasks from the 187 bug fixes that change a test
  • Long-context coding agents: one repository with nine years of history and 154 people

Before delivery

  • Any third-party engine or SDK code inside the repository is listed and left out before delivery.
  • The studio's parent company, which owns the code, confirms its rights in all of it before samples and terms.

Lot CODE-TENDER

An AI tender-response product's code, 2024 to 2026: Python and TypeScript with 1,465 commits and 272 merged pull requests

A B2B product that drafts tender responses with AI: a Python and FastAPI back end, a Next.js web app and a Word add-in, with 18 people in its history and tests at a 17% ratio. Repo Check rates it strong for building graded training tasks.

Original lines by Repo Check, including retained starter code the owner did not write
42,229
Commits
1,465
Merged pull requests
272
People in the repository's history
18
First and last commits
July 2024 to October 2026

What is inside

  • The back end, the web app, a Word add-in, document processing and AI generation, in one repository.
  • Test code at a 17% ratio to product code; Repo Check rates the repository strong for training tasks.
  • Measured by Repo Check 2.2.0 on 19 October 2026.
  • Disclosed: the repository commits 31 AI coding-assistant configuration files, and the owner estimates that about 80% of the code was written by people and 20% generated by AI.

Best for

  • Graded training tasks from a working B2B AI product, once its tests run offline
  • Coding agents on a current Python and TypeScript stack with a Word add-in

Before delivery

  • The owner reports customer documents, credentials and personal data scrubbed from the whole history; the transformed rows left in the database dumps are replaced with synthetic rows or removed, and the credential rotation finished, before any copy goes to a buyer.
  • The MIT notices for the starter code and the Office add-in template files stay with the code.
  • The rights of everyone in the history are confirmed before signing.

Lot CODE-AGENCY

99 private repositories from one software agency, with history from 2014: 7.6 million original lines in PHP, JavaScript and TypeScript

An agency's code with its full git history: PHP, JavaScript, TypeScript, Python and Vue, with 3 to 42 people in each repository's history and tests in 91 of the 99. All but one pass Repo Check's four gates, and 14 are rated strong for building graded training tasks.

Private repositories with full git history, 98 passing Repo Check's four gates
99
Original lines (generated and vendored code not counted)
7.6M
Commits
More than58,800
Merged pull requests, estimated from the commit history
More than30,480
First and last commits
2014 to 2026
People in each repository's history
3 to 42

What is inside

  • By language: PHP 3.7 million, JavaScript 2.8 million, TypeScript about 470,000, Python about 250,000, Vue about 162,000 lines.
  • Fifteen of the repositories grew from shared first commits (a group of 11 and two pairs), so their shared early history is counted once in the commit figure.
  • Tests in 91 of 99 (3,132 test files; Cypress, JUnit, Jest, Mocha, PHPUnit, Python tests, pytest), and CI on AppVeyor, CircleCI, GitHub Actions, GitLab CI, Travis CI.
  • At least 14,730 bug-fix commits, at least 514 of them changing a test (the shared early history counted once).
  • Repo Check rates 14 of them strong for graded training tasks, with at least 354 bug fixes that change a test; 13 of those pass all four gates.
  • For 96 of them the owner has stated, repository by repository, that the code was never public, is not a fork, and is the agency's own or held with its client's written consent to license it; the other 3 get the same statement before samples.
  • Fifteen were started in 2023 or later; in 76 of the other 84, 10% or less of the added lines date from 2023 or later.
  • Disclosed: 3 of the 99 repositories commit AI coding-assistant configuration files.
  • Measured by Repo Check 2.4.1 on 2026-10-26, 2026-10-27, repository by repository.

Best for

  • Graded training tasks: 14 repositories rated strong, with at least 354 bug fixes that change a test
  • Code-change and review models: more than 30,480 merged pull requests, estimated from the commit history
  • Coding agents on PHP, JavaScript and TypeScript code with its history

Before delivery

  • Credentials that Repo Check found in 53 of the repositories are rotated and removed before delivery.
  • Personal data in the code, such as email addresses, is removed before delivery.
  • Repo Check flagged at least 105 files with GPL-family license headers and 198 package manifests that could be published to public registries: the owner confirms which parts, if any, are published, and those parts and the GPL-licensed files are left out before samples.
  • Four repositories carry an open-source license file (two GPL, two MIT): the owner confirms what each covers, and GPL-licensed code is left out.
  • The agency confirms, repository by repository, which code is its own and which it built for clients, before samples and terms; client work is shown only with the client's written consent, or left out.

Lot CODE-JS-JAVA

82 private repositories from one owner, 2014 to 2023: about 922,000 original lines in JavaScript, Java, Kotlin and Objective-C

Code with its full git history from 2014 to 2023, most of it JavaScript and Java, with Kotlin, Objective-C and Swift repositories among them: 1 to 29 people in each repository's history and tests in 34 of the 82. Nine pass Repo Check's four gates; the other 73 miss at least one and are priced on request.

Private repositories with full git history, 9 passing Repo Check's four gates
82
Original lines (generated and vendored code not counted)
About922,000
Commits
More than19,900
Merged pull requests, estimated from the commit history
More than4,720
First and last commits
2014 to 2023
People in each repository's history
1 to 29

What is inside

  • By language: JavaScript about 661,000, Java about 136,000, Kotlin about 54,000, Objective-C about 37,000, Swift about 14,000 lines.
  • Four of the repositories grew from shared first commits (two pairs), so their shared early history is counted once in the commit figure.
  • Tests in 34 of 82 (96 test files; JUnit, Jest, Mocha), and CI on GitLab CI, Travis CI.
  • At least 2,596 bug-fix commits, at least 6 of them changing a test (the shared early history counted once).
  • By main language: 44 repositories in JavaScript, 24 in Java, 8 in Objective-C, 4 in Kotlin, and one each in Swift and C#.
  • Nine pass all four of Repo Check's standing-rate gates (3 core contributors, 100 commits, 10 merged pull requests and 10,000 original lines); the other 73 miss at least one, and 22 of the 82 have 10,000 original lines or more.
  • For all 82 the owner has stated, repository by repository, that the code was never public, is not a fork, and is the owner's own or held with its client's written consent to license it.
  • The newest commit in the set dates from 2023, and in every repository under 1.5% of the added lines date from 2023 or later.
  • Measured by Repo Check 2.4.1 on 2026-10-26, repository by repository.

Best for

  • Code written before 2023: in every repository, under 1.5% of the added lines date from 2023 or later
  • Coding agents on JavaScript and Java code with its history, and on Kotlin, Objective-C and Swift
  • Repair data: at least 2,596 bug-fix commits

Before delivery

  • Credentials that Repo Check found in 34 of the repositories are rotated and removed before delivery.
  • Personal data in the code, such as email addresses, is removed before delivery.
  • Repo Check flagged 10 files with GPL-family license headers and 31 package manifests that could be published to public registries: the owner confirms which parts, if any, are published, and those parts and the GPL-licensed files are left out before samples.
  • The owner confirms, repository by repository, which code is the owner's own and which was built for clients, before samples and terms; client work is shown only with the client's written consent, or left out.

Lot CODE-RAILS-SET

16 private codebases from one software company, 2008 to 2026: mostly Ruby on Rails, with more than 57,000 commits

Applications and business systems built across several stacks, with their git histories: Rails in 14 of the 16, React or Next.js in 8, and 6 to 101 people in each codebase's history. The figures are the owner's own analysis; Repo Check measures each repository before samples.

Codebases with git history intact, by the owner's analysis
16
Commits, by the owner's analysis (a codebase kept in two copies is counted once)
More than57,000
From the earliest first commit to the latest commit, by the owner's analysis
2008 to 2026
People in each codebase's history, by the owner's analysis
6 to 101
Codebases with more than five years between first and last commit, by the owner's analysis
7 of 16

What is inside

  • By the owner's analysis: Rails in 14 of the 16, React in 6 and Next.js in 5; Ruby and JavaScript are the main languages, with TypeScript in 3, Python in 2, and Java and Swift in one. One of the 16 is a game-engine project (GDScript and scene files) that holds 1.2 million of the counted lines; another is a JavaScript app with Java and Swift native code.
  • Six of the 16 have pull requests in the export, 702 merged in all; for three more the export reached pull request data and found none, and for the other seven it has none.
  • The owner's figures, read against Repo Check's four standing-rate thresholds (3 core contributors, 100 commits, 10 merged pull requests and 10,000 original lines), put six of the 16 above all four and the other 10 short only on merged pull requests, seven of them because the export has no pull request data. The export counts every contributor and every authored line, not core contributors and original lines, and Repo Check reads merged pull requests from the commit history itself, so its run decides.
  • The owner's analysis counts about 7.4 million authored lines, but 5.9 million of them sit in two codebases with 1,563 and 804 commits, where the same export counts under one token of code per line, so most of those lines are likely data or generated files rather than source; the other 14 hold about 1.5 million lines. Repo Check's original-line count, which leaves out vendored and generated code, comes with its run.
  • Five of the 16 had commits in 2025 or 2026, and seven have more than five years between their first and last commits.
  • One codebase is kept in two copies with the same first and last commit days; it is counted once, by the larger copy.
  • Databases in the code: Redis in 12, PostgreSQL in 7, MySQL or MariaDB in 5.
  • The owner describes the set as private repositories, legacy codebases among them, holding both the company's own code and code written for its clients.

Best for

  • Coding agents and code models for long-lived Ruby on Rails applications
  • Full-stack web work: Rails back ends with React or Next.js in 8 of the 16
  • Legacy code with long histories: seven codebases with more than five years between first and last commit

Before delivery

  • Non-exclusive licenses, as the owner prefers: the company keeps ownership of its code.
  • Repo Check runs on each repository before samples and terms; its report sets the figures and the price.
  • The owner confirms, repository by repository, that the code is the company's own or held with its client's written consent to license it, was never public and is not a fork, before samples; client work is shown only with the client's written consent, or left out.
  • The owner's analysis flags items for security review in 15 of the 16; any credentials among them are rotated and removed before delivery.
  • Personal data in the code, such as email addresses, and the database dump files committed to two of the repositories are removed before delivery.
  • One repository carries a license file: the owner confirms what it covers, and any open-source-licensed code is left out before samples.

Lot CODE-TRADING

One private Python trading framework from a sole developer: strategy replay, risk controls and broker execution for FX and CFD

A systematic trading codebase in Python with strategy replay, risk controls, order execution through a broker connection, persistent trade state and broker reconciliation. The current code is in a private git repository with a short history; earlier work is kept as dated source snapshots, patches, engineering reports and test records. The figures are the owner's own description; Repo Check measures the repository before samples.

Private git repository, with a short history (owner's words)
1
Developer: one person wrote the code (owner's description)
1
Source licenses or IP rights granted to anyone so far (owner's words)
None

What is inside

  • Strategy replay, risk controls, FX and CFD execution through a broker connection, persistent trade state and broker reconciliation, by the owner's description.
  • Beside the git repository: dated source snapshots, correction requests, patches, engineering reports, test records and verification records, by the owner's description.
  • Pre-revenue; one developer. Repo Check reports lines, commits and authorship before samples.

Best for

  • Coding agents for trading, risk and order-management systems in Python
  • Engineering-process data: code together with its correction requests, patches and test records

Before delivery

  • Non-exclusive licenses; the owner keeps the code.
  • Repo Check runs on the repository before samples and terms; its report sets the figures and the price.
  • The owner confirms, before samples, that the code is their own, was never public and has not been licensed to anyone.
  • Historical market data used by the replay is left out unless its license allows sharing it.
  • Any credentials, broker keys and account details are removed before delivery.

Lot CODE-STRATEGY-SET

32 private repositories of automated trading code in MQL4, MQL5 and Python, written in 2026

One trader's own strategy and automation code for a retail trading platform: 32 private repositories with more than 320 files, some codebases over 2,500 lines, never public by the owner's account. The owner began writing them this year with the help of AI tools. The figures are the owner's; Repo Check measures the Python code before samples, and does not read MQL.

Private repositories (owner's count)
32
Files across the repositories (owner's count)
More than320
Lines in some of the codebases (owner's words)
Over2,500

What is inside

  • Languages: MQL4, MQL5 and Python, run through a retail trading platform, by the owner's description.
  • Never public, by the owner's account; the owner offers read-only access to one repository as a sample.
  • Written in 2026 by one author with the help of AI tools, by the owner's account; Repo Check reports authorship and history.

Best for

  • Code models for trading-platform scripting languages that public code covers thinly
  • Coding agents for strategy automation and order logic

Before delivery

  • Non-exclusive licenses; the owner keeps the code.
  • Repo Check runs on each repository before samples and terms; its report sets the figures and the price of the Python code, and the MQL code is priced on request.
  • The owner confirms, before samples, that the code is their own, not written for an employer, a trading firm or a client, and holds no third-party indicator or library code.
  • Trading account credentials, broker keys and API keys are removed before delivery.

Lot CODE-JOBS

Code from a job and career platform company, 2024 to 2026: one PHP and JavaScript monorepo with 2.7 million original lines

A monorepo from a job and career platform company: PHP and JavaScript with some Svelte and TypeScript, 9 people in its history and 751 bug-fix commits since 2024. Too large for the standing rate, so it is priced on review.

Monorepo with full git history, passing Repo Check's four gates
1
Original lines (generated and vendored code not counted)
2.7M
Commits
4,019
Merged pull requests, estimated from the commit history
About74
First and last commits
2024 to 2026
People in the repository's history
9

What is inside

  • By language: PHP 1.8 million, JavaScript about 968,000, Svelte about 4,000, TypeScript about 4,000 lines.
  • 53 test files (3,346 test lines; Jest, PHPUnit, Playwright), and CI on CircleCI, GitHub Actions, Travis CI.
  • 751 bug-fix commits.
  • Measured by Repo Check 2.3.0 on 2026-10-22, repository by repository.
  • Disclosed: the repository commits 8 AI coding-assistant configuration files.

Best for

  • Coding agents on a large PHP and JavaScript codebase
  • Repair data: 751 bug-fix commits from 2024 to 2026
  • Long-context code models: 2.7 million original lines in one repository

Before delivery

  • Credentials that Repo Check found in the code are rotated and removed before delivery.
  • Repo Check flagged 37 files with GPL-family license headers and 25 package manifests that could be published to public registries: the owner confirms which parts, if any, are published, and those parts and the GPL-licensed files are left out before samples.
  • Any third-party code inside the repository is listed and left out before delivery, and the line count is restated after that.
  • Personal data in the code, such as email addresses, is removed before delivery.
  • The company confirms its rights in the code before samples and terms.

Lot CODE-POOL

Private code from 23 independent teams and small software companies: more than 240 repositories

Codebases their owners describe as never public, from founders, agencies and small product companies: TypeScript, Python, Go, Java, Ruby, Swift, Flutter and more. Each repository is measured by Repo Check, including its never-public check, and priced on its own before samples, so a buyer can take the whole pool or pick by stack.

Owners in the pool
23
Repositories, by the owners' own counts and Repo Check
More than240
Measured by Repo Check in the owner portal (more reports came by email)
17

What is inside

  • 4 repositories in TypeScript, Swift and JavaScript, active 2021 to 2023: about 80,000 lines, 3,103 commits, about 537 merged pull requests, tests in all 4.
  • 13 repositories, mostly TypeScript, JavaScript and Python, 2014 to 2026: about 401,000 lines and 3,680 commits; 3 priced at the standing rate today.
  • 10 TypeScript repositories, over 200,000 lines, over 1,000 commits and over 100 merged pull requests (owner's description).
  • 107 GitLab repositories in Ruby, Java, JavaScript, SQL and Python (owner's description).
  • 62 Go repositories with 100 to 1,000 commits each (owner's description).
  • 17 repositories, 1 to 3 years of history, 1 or 2 people each (owner's description).
  • 30 repositories by the owner's own audit, AI-assisted, which the owner is screening for agent, workflow and automation work: 2 sent and measured by Repo Check so far, the only 2 counted here (owner's description).
  • A software firm's own SaaS products: GitLab repositories in Laravel, Node.js, React, TypeScript and Python, over 100 merged pull requests (owner's description).
  • Repositories in TypeScript, Python, React, Android and Go, 100 to 1,000 commits each (owner's description).
  • Repositories in Flutter, Dart and Python, 100 to 1,000 commits each (owner's description).
  • Python repositories (owner's description), among them a financial reconciliation system that Repo Check measured at about 122,700 original lines, 227 commits and 2 people.
  • GitHub repositories, size to be measured: the owner's two forms disagree (owner's description).
  • A multi-tenant messaging-automation SaaS in Python, TypeScript and Next.js, over 100 merged pull requests (owner's description).
  • One TypeScript repository with over 100 merged pull requests (owner's description).
  • One TypeScript and JavaScript product repository (owner's description).
  • TypeScript repositories with 10 to 100 merged pull requests (owner's description).
  • 12 Bitbucket repositories, a first selection from more than 100 across about 10 years of production code (owner's description).
  • TypeScript repositories on GitLab with 10 to 100 merged pull requests, also licensed elsewhere non-exclusively (owner's description).
  • 7 PHP and Laravel backend repositories of a hospital information system, its billing, lab, clinical, pharmacy and other modules: 50,000 to 200,000 lines, over 1,000 commits and over 100 merged pull requests (owner's description; checked for patient data before any sample).
  • 6 repositories in TypeScript, JavaScript, Python and Java, over 1,000 commits and 10 to 100 merged pull requests (owner's description).
  • One TypeScript, React and Node.js repository: 10,000 to 50,000 lines, 1 to 3 years of history, tests in CI (owner's description).
  • One of a game studio's titles, and possibly some more of them (owner's description).
  • Two game projects: a multiplayer Unreal Engine project with full source, mostly C++, built around dedicated servers, and a Unity mobile game for Android and iOS on a PlayFab backend (owner's description).

Best for

  • Coding-agent training across many stacks and team sizes
  • Buyers who pick repositories by language, size and history before licensing

Before delivery

  • Every repository is measured by Repo Check before samples; figures the owners gave are replaced by measured ones.
  • Third-party, GPL-licensed and already published code is listed and left out, and personal data is removed, before delivery.
  • Each owner confirms their rights in their own code before terms.

Lot CODE-RAILS19

A video-testimonial product's code, 2019 to 2021: one Rails and React codebase with 4,367 commits on its default branch, measured by Repo Check

A small company's web product for collecting, processing and publishing video testimonials, with referral features, offered as its fixed historical code with its history. Buyers see the owner's summary and Repo Check's aggregate figures only; the owner's rights documentation is in progress.

Commits on the default branch, by Repo Check (3,272 plus 1,095 merge commits)
4,367
Lines of original code, by Repo Check (12,639 code and 2,581 test lines)
15,220
Merged pull requests, estimated by Repo Check from commit messages
817
Contributor identities in git, bots left out (git identities, not verified people)
6
Historical repair candidates with isolated before and after checks, by the owner's count: the repair-task route, shown apart from the license
25
Public: the repository was always private, by the owner's account
Never

What is inside

  • A Rails and React web product for collecting, processing and publishing video testimonials, with referral features.
  • Languages by Repo Check: Ruby 56%, JavaScript 44%. Test frameworks: RSpec and Jest; 56 test files, with test lines at 20% of code lines.
  • History by Repo Check: first commit in 2019, about 1.9 years of active history over 24 active months; 258 bug-fix commits, 24 of them with test changes. The most active identity made 82% of the commits. Contributor counts are git identities with bots left out, not verified people.
  • Code license: all four standing-rate gates met (contributors, commits, merged pull requests and original lines). Repo Check rates it "possible" for training-task buyers: tests, a mainstream stack and bug fixes that changed tests pass; no continuous integration was found, and the active history is 1.9 years, under the 2 years the check asks for, so no task value is priced.
  • Repair-task route, apart from the code license: the owner reports 25 historical repair candidates with isolated checks that fail before the fix and pass after it, 23 in Ruby and 2 in JavaScript, and seven fixes that added or changed original regression tests. The isolated before/after checks use retrospective harnesses with synthetic inputs and lightweight stand-ins; the original test suites have not been verified in CI or a clean container. In the owner's words, these counts "size the existing curation pipeline"; they are not finished, portable buyer tasks. Priced on review, apart from the license.
  • Open-source libraries come in only as declared dependencies; no open-source or vendor code is copied into the repository (owner's description).
  • The full application build has not been verified yet.

Best for

  • Code-change and repair models on a Rails and React product with its git history
  • Repair tasks built from historical bug fixes with isolated before and after checks (priced apart)

Before delivery

  • Non-exclusive licenses only, as the owner asks.
  • The owner's rights documentation is in progress: any license or binding commitment needs it completed.
  • Samples and final terms remain subject to the owner's written approval.
  • Buyers see the summary and aggregate figures only. Source code, snippets, repository access, code evaluation and training are on hold.
  • The code license and the repair-task route are shown, priced and agreed separately.
  • Repo Check found 4 possible credentials in the current code and 4 distinct possible credentials in the history, and 18 email addresses in the code: credentials are rotated and removed, and personal data removed, before delivery.

Left out

  • Customer records, recordings and testimonials

Lot CODE-RAILS25

An operating business's production Rails app, August 2025 to October 2026: 11,268 commits, 4,665 of them with no AI co-author

The codebase behind a live business, built by its owner with substantial AI assistance. Git credits 4,665 commits to the owner alone and 6,454 to the owner with an AI co-author, so work with and without an AI co-author sits side by side in one history. Repo Check measured 393,465 lines of original code, 167,262 of them tests.

Commits, August 2025 to October 2026: the owner's count from git, and Repo Check's 10,265 plus 1,003 merge commits on the default branch
11,268
Commits with an AI co-author tag, by the owner's count from git
6,454
Lines of original code, by Repo Check (226,203 code and 167,262 test lines)
393,465
Merged pull requests, estimated by Repo Check from commit messages
793
Historical repair behaviors across three fix commits, by the owner's count: the repair-task route, shown apart from the license
4

What is inside

  • A production Rails app; the owner directed and reviewed development with substantial AI assistance (owner's description).
  • Languages by Repo Check: Ruby 97%, JavaScript 2%. Tests: RSpec, 1,234 test files, with test lines at 74% of code lines.
  • History by Repo Check: first commit in 2025, last activity 14 days before the check, 15 active months; 2,207 bug-fix commits, 729 of them with test changes. 3 contributor identities, the most active with 99% of the commits; the owner counts 4,665 commits credited to the owner alone. Contributor counts are git identities with bots left out, not verified people.
  • Disclosures by Repo Check: 50 AI-assistant settings files are committed in the repository; password-like strings and email addresses in test files read as fixtures.
  • Code license: all four standing-rate gates met, and the standing-rate figure is above the standard range of $150 to $1,500 per repository, so it is priced on review. Repo Check rates it "possible" for training-task buyers: tests, a mainstream stack and bug fixes that changed tests pass; no continuous integration was found, and the active history is 1.2 years, under the 2 years the check asks for, so no task value is priced.
  • Repair-task route, apart from the code license: the owner reports four historical repair behaviors across three fix commits, with isolated checks improving from 32 of 44 passing before to 44 of 44 after. The isolated before/after checks use retrospective harnesses with synthetic inputs and lightweight stand-ins; the original test suites have not been verified in CI or a clean container. In the owner's words, these counts "size the existing curation pipeline"; they are not finished, portable buyer tasks. Priced on review, apart from the license.

Best for

  • Research on AI-assisted development: one production app's history, with commits made with and without an AI co-author
  • Coding-agent training on the 4,665 commits credited to the owner alone
  • Repair tasks from historical fixes with isolated before and after checks (priced apart)

Before delivery

  • Non-exclusive licenses only, as the owner asks.
  • Samples and final terms remain subject to the owner's written approval.
  • Buyers see the summary and aggregate figures only until the owner approves samples.
  • The code license and the repair-task route are shown, priced and agreed separately.
  • Repo Check found 5 provider-format keys in test files and 3 distinct possible credentials in the history, and 39 email addresses in the code: credentials are rotated and removed, and personal data removed, before delivery.

Lot CODE-TS-JS

47 private repositories from one owner, 2019 to 2026: about 962,000 original lines in TypeScript, JavaScript and Kotlin

Code with its full git history from 2019 to 2026, most of it TypeScript, with JavaScript, Kotlin, Python and Go repositories among them: 1 to 68 people in each repository's history and tests in 26 of the 47. Eighteen pass Repo Check's four gates; the other 29 miss at least one and are priced on request.

Private repositories with full git history, 18 passing Repo Check's four gates
47
Original lines (generated and vendored code not counted)
About962,000
Commits
More than46,400
Merged pull requests, estimated from the commit history
More than25,600
First and last commits
2019 to 2026
People in each repository's history
1 to 68

What is inside

  • By language: TypeScript about 683,000, JavaScript about 188,000, Kotlin about 77,000, Python about 8,000, Go about 5,000 lines.
  • Two of the repositories grew from the same first commits, so their shared early history is counted once in the commit figure.
  • Tests in 26 of 47 (612 test files; JUnit, Jest, Playwright, Python tests, Vitest, pytest), and CI on GitHub Actions, GitLab CI.
  • At least 20,738 bug-fix commits, at least 499 of them changing a test (the shared early history counted once).
  • By main language: 35 repositories in TypeScript, 5 in JavaScript, 4 in Python, one each in Kotlin and Go, and one empty repository (one commit, no source code).
  • Eighteen pass all four of Repo Check's standing-rate gates (3 core contributors, 100 commits, 10 merged pull requests and 10,000 original lines); the other 29 miss at least one, and 20 of the 47 have 10,000 original lines or more.
  • Most of it is recent: 29 of the 47 were started in 2023 or later, and in 11 of the other 18 more than half of the added lines date from before 2023.
  • For all 47 the owner has stated, repository by repository, that the code was never public, is not a fork, and is the owner's own or held with its client's written consent to license it.
  • Measured by Repo Check 2.4.1 on 2026-10-28, repository by repository.

Best for

  • Coding agents on TypeScript and JavaScript code with its history, with Kotlin, Python and Go besides
  • Repair data: at least 20,738 bug-fix commits, at least 499 of them changing a test
  • Current TypeScript practice: most of the code was written from 2023 on

Before delivery

  • Credentials that Repo Check found in 14 of the repositories, and environment files in 12, are rotated and removed before delivery.
  • Personal data in the code, such as email addresses, is removed before delivery.
  • Repo Check flagged 43 package manifests that could be published to public registries: the owner confirms which parts, if any, are published, and those parts are left out before samples.
  • The owner confirms, repository by repository, which code is the owner's own and which was built for clients, before samples and terms; client work is shown only with the client's written consent, or left out.

Lot CODE-MOBILE-GAMES

A game studio's older catalog: about 30 games from 16 years of work, mostly Unity and mostly mobile

The source code of a studio's older titles, built over 16 years, mostly in Unity for mobile, with some PC and VR games among them. The studio is sizing the catalog and the third-party content to strip out before samples.

Older games, by the studio's first estimate
About30
Of the studio's development
16 years
Most of the games; mostly mobile, some PC and VR
Unity

Best for

  • Game-building agents on Unity code
  • Code models for mobile games, with some PC and VR

Before delivery

  • Third-party content is stripped out, and the studio confirms each game's code was never public, before samples.

Lot CODE-KIDS-GAMES

A game studio's code: a repository per game, around 17 of them on the studio's own engine

Separate repositories for each of the studio's games, around 17 of which share its own core engine. Sizes and history are measured before samples.

Game repositories on the studio's shared engine
Around17
Each game in its own repository
1 per game

Best for

  • Game-building agents on a shared engine and the games built on it
  • Code models for game development

Before delivery

  • Every repository is measured by Repo Check, and the studio confirms the code was never public, before samples.
  • Only the studio's own code and assets are licensed: player data, and anything players make in the games, stay out.

Lot CODE-DELPHI

A software house's Delphi and Pascal enterprise systems, 2000 to 2021: hundreds of modules and millions of lines, written without AI

Systems in real-world use for years, from municipal and urban planning to association and charity management and national-scale public services, with the team's own IDE and SQL-based database. The version history lives in the team's own tracking system, kept since the early 2000s, not in git.

Built and maintained without AI tools
2000 to 2021
Lines of code, by the owner
Millions
Of modules
Hundreds

Best for

  • Code models for Delphi and Pascal
  • Legacy modernization and migration agents

Before delivery

  • The version history comes from the owner's own tracking system; its export format is agreed before samples.
  • The code only: no data hosted in or processed by the systems, whether users', patients' or organizations', is included.

Lot CODE-UNITY-GAME

One large Unity game in early access: its source with private backups since 2021, from before AI assistance

The source of a commercial Unity game in development since 2020, with dated private backups of its versions from 2021 on, never public. The studio added AI support only in 2025, so the earlier versions were written without it.

In development; dated backups from 2021
Since 2020
Versions written without AI support
Before 2025

Best for

  • Game-building agents on a large Unity codebase
  • How one codebase grew across dated versions, all written without AI support

Before delivery

  • Third-party assets and plugins inside the project are listed and left out before samples.

Lot CODE-STARTUP4

Four large private codebases from a software company's social, AI and developer-tool products

Production code with tests running in CI, written by the company's engineering team over multiple years. The owner puts it at millions of lines in all, with its engineering records; Repo Check measures history, contributors and reviews, and has not run yet.

Private repositories, extra-large on the owner's form
4
Tests, on the owner's form
In CI

What is inside

  • Hosted on GitHub and GitLab, never public, and not licensed to anyone else, on the owner's form.

Best for

  • Coding agents on large production codebases
  • Code models for social, AI and developer-tool products

Before delivery

  • Any open-source code inside the repositories is identified and left out before samples (the owner is not sure there is any).

Lot CODE-UNREAL-GAME

The source of one finished PC game, built in Unreal Engine 5 and released on Steam in 2024

A finished commercial game, made and published by one studio and on sale for Windows since 2024. The studio confirms before samples whether the project and editor files, the version history and a build that opens come with the source.

The game's engine
Unreal Engine 5
Released on Steam, for Windows
2024

Best for

  • Game-building agents on a finished Unreal Engine 5 game
  • Code models trained on a shipped game's codebase rather than snippets

Before delivery

  • The engine's own source and any licensed third-party art, assets and plugins stay out; the studio lists them before samples.

7 lots

More data

Books, CAD and design files, and other data

Lot TXT-FICTION

A self-published fiction backlist: 69 English titles and 32 translations in thriller, crime, romance, science fiction and fantasy

One author-publisher's catalog of published novels: 69 English titles and 32 translations into five languages, listed edition by edition with the year and, for most, the word count in the owner's title lists.

English titles, published 2016 to 2024, by the owner's title list
69
Translations (German 18, French 9, Italian 3, Spanish 1, Dutch 1), published 2020 to 2024
32
Words, counted on 94 of the 101 editions by the owner's lists
About4.6M
Genres
5

What is inside

  • By the owner's lists: 101 editions, 69 English originals with about 3.2 million words (word counts on 65 of them, 10,000 to 125,000 each) and 32 translations with about 1.3 million words (word counts on 29 of them).
  • Every edition on the lists is held as EPUB and Word files, marked as the owner's own, and marked as not edited or translated by an AI tool.
  • 14 of the 32 translations name an English title found on the English list, so those pair with their originals.

Best for

  • Long-form fiction for language models
  • Multilingual parallel editions: at least 14 translations pair with their English originals

Before delivery

  • Translators' and co-authors' rights are confirmed title by title when a buyer shows interest, before samples.

Lot TXT-BACKLIST

An Indian publisher's backlist of 808 titles, listed with ISBN, author and year

808 academic and reference titles across 22 subjects, published from 2004 to 2026 and all held in PDF; the publisher states it holds complete digital rights to every title.

Titles by distinct ISBN, with author and year
808
Subjects; some titles are listed under more than one
22
Every title, in the publisher's words
PDF

What is inside

  • The largest subjects are business and management, mechanical engineering, computer science and IT, botany, civil engineering and biotechnology; others include earth and environmental science, electronics, mathematics, physics, chemistry, agriculture, medicine and pharmacy.
  • Most titles were published between 2008 and 2023. The list gives ISBN, author, title and year for each, in 1,153 listings, because a title can sit under more than one subject.

Best for

  • Long-form published text for language models
  • Licensed book content for AI answers

Before delivery

  • Titles whose author contracts predate AI are licensed only with the author's written consent; freelance material and agency photos stay out.

Lot TXT-SAT

More than 6,000 original SAT questions, each with a detailed explanation

A test-prep company's own SAT bank in the current digital format, covering both sections, Reading and Writing, and Math: every question written by its staff teachers, each with a detailed explanation.

Original SAT questions, each with an explanation
More than6,000
Students who attempted the questions
Thousands

What is inside

  • Both sections, Reading and Writing, and Math, in the current digital SAT format; none adapted from official papers or released tests.
  • Results per question by student score band and actual SAT result, aggregate only (to be confirmed), for example the share of top-band students who answered correctly. No student information of any kind is shared.
  • The teachers' employment terms assign their work to the company, and the agreements are ready for review. No edit history is kept.
  • The company also holds about 2,000 ACT questions, and material for other curricula and tests not yet sized.

Best for

  • AI tutors and test-prep assistants
  • Question generation calibrated to real student difficulty
  • Evaluating written explanations in math and reading

Before delivery

  • The teacher agreements are reviewed under NDA before samples; student records never leave the owner.
  • The aggregate results go in only where the owner's student terms allow that use.

Lot REVIEW-AGENT

181 runs of a code-review agent on live pull requests in public repositories, 18 of them scored against what the human reviewers asked for

Full trajectories of an AI agent reviewing real pull requests in 14 public repositories while they were open: every tool call against an index of the repository (callers, dependents and source lines) and the review it wrote. Labeled reviews add an answer key built from the human reviewers' comments and a score for each ask.

Completed reviews, made only with open-weights models
181
Pull requests in 14 repositories
About178
Further reviews a week the owner says it can add, made with other open-source review tools
About300

What is inside

  • Each run holds the pull request under review, every event in order (each tool call with its arguments, each result preview, findings as they were raised, token use per step) and the final review: findings with path, lines, claim, evidence and severity.
  • A labeled review adds a link to the pull request and its review thread, an answer key of the distinct asks the human reviewers made, and a hit, partial or miss verdict for each ask, with recall and precision. 18 reviews are labeled so far.
  • A model wrote each answer key from the human comments alone, author and bots removed, without seeing the agent's findings; a second model call gave each verdict, and recall and precision are computed in code.
  • Also offered: a hosted environment where a buyer's agent reviews a pull request through the same repository-index tools and is scored on the owner's side, so the answer keys never leave it. Tasks can be drawn from pull requests opened after a model's training cutoff, so they were not public when it trained.

Best for

  • Code-review agents
  • RL environments with held-out answer keys
  • Tool-use trajectories for agent training

Before delivery

  • Each model used in the runs and labels is listed with its license and checked for this use before signing.
  • The human reviewers' comments are linked, not copied.
  • The reviewed repositories' licenses are listed, and their notices travel with the excerpts the trajectories quote.

Lot FUNNEL-AGG

One business's marketing results in aggregate, December 2024 to August 2026: email and SMS campaign summaries, funnel summaries and web channel totals

Performance summaries from an operating business's email and SMS platform, its product-analytics funnels and its web analytics, in aggregate only. Customer records, message content, user profiles, session replays and ad-platform results are left out.

December 2024 to August 2026, the period proposed; some sources' histories run later
21 months
Sources: email and SMS campaigns and flows, product-analytics funnels, web channel and funnel totals
3
Personal data: aggregates only
None

What is inside

  • Email and SMS campaign and flow performance summaries.
  • Aggregate funnel summaries from the business's product analytics.
  • Website channel and funnel totals, subject to a current report.
  • A screened export is not assembled yet; the owner keeps an inventory of the sources, periods and detail levels.

Best for

  • Environments and evaluations for marketing-analytics agents, built on one real business's results

Before delivery

  • Non-exclusive licenses only, as the owner asks.
  • Aggregate figures only: nothing at the level of a person.

Left out

  • Customer records, message content, user profiles and session replays
  • Results from the ad platform

Lot TXT-ACADEMIC

More than 1.1 million titles, at least 95% academic, in seven main languages, across STEM and other fields, held as PDF

An academic publisher's catalog, at least 95% academic titles across essentially every STEM field and more, 125 pages on average and author-original by the publisher's policy: its author agreements bar AI-generated submissions and it scans for them, which it says is no guarantee. It holds non-exclusive rights that include licensing.

Titles in seven main languages
More than1.1M
In English, and 120,000 to 170,000 in each of the six others
About250,000
Average length; PDF files
125 pages

What is inside

  • By language: about 250,000 titles in English and 120,000 to 170,000 in each of German, Spanish, French, Portuguese, Russian and Italian, with smaller sets in many other languages.

Best for

  • Long-form academic text in many languages
  • Multilingual STEM language models

Before delivery

  • The publisher's author-agreement versions, and which titles each covers, are reviewed at the level of the agreement before samples, not title by title; the publisher's warranty in the license covers the titles in scope, and titles no agreement covers are left out.

Lot TXT-PRACTICE

A professional environmental institute's practice guides, with its position statements and policy submissions

Guidance written by the institute's expert volunteers and staff and published under its name: practice guides on impact assessment and environmental accounting. The institute owns the IP in each.

Four put forward first by the institute
Practice guides
And policy submissions, also offered
Position statements

What is inside

  • Volunteers sign nothing, but every publication carries the institute as author; it offers a statutory declaration that those credited are part of it.
  • The guides are on the institute's public website, which the institute says gives no permission to use them.

Best for

  • Environmental and impact-assessment assistants
  • Expert professional guidance text

Before delivery

  • The institute confirms its ownership of each publication in writing at signing, with the statutory declaration it offered; figures, maps, photos or tables reproduced from other publishers stay out unless their owners agree.

2 lots

Training tasks

Graded tasks built from real fixes and their tests

Lot 02-B-TASKS

SWE tasks from 233 real fixes that changed a test, with an evaluation split held back for one buyer

Each candidate is a real repository at a real commit, the change a developer made to fix it and the test that changed with it: the shape of SWE-bench-style tasks, where a grader runs the tests. Tasks are checked in a clean container before delivery, and a split can be kept out of every training license as one buyer's private evaluation set.

Candidate tasks: fix commits that changed a test
233
Source repositories, CI configured in all 14
14
Commits marked as fixes in the same repositories
5,600+
Candidates in the largest repository (more than 2 million lines, 155 people)
131

What is inside

  • A candidate is a commit whose subject marks a fix (fix, bug, patch, hotfix, revert, resolve, correct, regression) and that changed both a test file and code.
  • A graded task is the repository at the commit before the fix, a task statement, the developer's change as the reference solution and the tests that grade it.
  • A task is delivered only if its tests fail before the fix and pass after it in a clean container.
  • Spread: 131 candidates in the largest repository, 22 in a TypeScript repository of more than 480,000 lines, 19 in a 79,000-line JavaScript repository; 2 of the 14 have none.
  • The owner reports 0 failing tests in the latest run of each suite; Rightmark has not seen the logs.
  • Disclosed: 204 of the 233 sit in repositories that commit AI coding-assistant configuration files, and 12 sit in the repository built for a client, held until the client consents.

Best for

  • RL training for coding agents on fail-to-pass bug-fix tasks
  • A held-out evaluation set on code that is not in public training data, after the never-public check
  • Grader and reward-model work with the developer's own fix as the reference
  • Comparing agent patches with the change a developer actually made, and its tests

Before delivery

  • The owner reports that every credential found in the code or its history has been rotated. Credential strings still in the code or its history are removed or shown to be dead before delivery.
  • Any repository found to have been public is left out.
  • The owner's rights are reviewed lot by lot.
  • Once you accept the NDA in Rightmark's buyer portal, Rightmark shares repository-level metrics (names withheld) and the detailed source list. Code samples come only with the owner's agreement.
  • The owner's name, product names and repository names are shared only when the owner agrees to an introduction.
  • Each task passes the clean-container check before delivery: its tests fail before the fix and pass after it.
  • The evaluation split is chosen with the buyer, delivered to that buyer only and kept out of every training license of these tasks.
  • Tasks from the client-built repository wait for the client's consent.
  • 5 of the 14 source repositories use private package registries; those packages ship with the tasks.

Left out

  • Candidates that fail the clean-container check
  • 12 candidates in the client-built repository, until the client consents
  • The full git history of the 14 repositories (that is lot B)

Lot MKT-TASKS

24 test-backed repositories with 1,052 fix-plus-test commits, for graded coding tasks

Every repository here carries its own test suite (5,254 test files and 613,518 lines of tests over 2,158,360 lines of mostly PHP and TypeScript), 18 have CI configuration, and 1,052 past fixes changed tests along with the code. Each such fix is a candidate task that the project's own tests can grade.

repositories with their own test suites (16 on PHPUnit, 18 with CI configuration)
24
test files (613,518 lines of tests)
5,254
fix commits that changed tests with the code, of 13,001 fixes
1,052
original lines (58.0% PHP, 37.7% TypeScript)
2,158,360
tests passed when one of the 24 was run offline (owner-reported; 38 failed)
2,538

What is inside

  • Main languages: 15 PHP, 7 TypeScript and 2 Python repositories. Frameworks found for the tests: PHPUnit (16 repositories), Jest (4), Vitest (4), Playwright (3) and pytest (2).
  • 15,351 merged pull or merge requests and 48,683 commits; 22 of the 24 have 10,000 lines or more, 21 have 10,000 lines or more and 3 or more contributors, and 11 have 50,000 lines or more.
  • Each repository passed a task-readiness screen: a test suite of at least 5 test files, with test lines at 10% or more of code lines, on a mainstream stack, plus at least two of CI configuration, 10 or more fixes that changed tests, and 2 or more active years with 3 or more core contributors.
  • Disclosed per repository: 18 of the 24 have half or more of their added lines from 2023 on, and 21 commit AI-assistant settings files.
  • Offline runs so far, reported by the owner on 24 October, cover 2 of the 24: a 67,830-line TypeScript repository passed 2,538 Jest tests and failed 38, run on public dependencies (7 of 323 suites failed to load), and a 6,589-line PHP repository passed 9 of 12 with platform requirements ignored, so that one is indicative.
  • 21 of the 24 had commits in the 90 days before measurement (21 October 2026).

Best for

  • SWE-bench-style task building: the issue, the failing test, the fix
  • RL environments for coding agents graded by each project's own suite
  • Regression-test generation from fixes that changed tests with the code
  • Agent evaluation on PHP and TypeScript product services

Before delivery

  • Priced per repository until each suite runs green offline; sold per graded task after each task passes a clean-container check (the reference fix passes, doing nothing fails).
  • The same title, never-public, secrets, copyleft and masking conditions as the code lot.
  • A per-repository disclosure of when the code was written and where AI-assistant settings are committed.
  • Evaluation samples restricted to evaluation, deletion on request.

Left out

  • Any repository whose suite does not run
  • The 2 test-backed repositories under 10,000 lines (offered with the code lot)
  • Everything held out of the code lot

2 lots

Bundles

Lots from more than one company, licensed in one order

Lot BND-2-CODE

Code from two B2B software companies: 404 repositories, more than 9.4 million lines, with tickets for 30 of them

Two stacks in one order: a sales-software company's code, about half of it C# (138 repositories; the main application runs about 12 years, to October 2026, with 193 contributors) and a second software company's JavaScript and TypeScript code (266 repositories, 30 of them delivered with their support tickets, engineering chat and work tracking). More than 206,000 commits and nearly 119,000 merged pull requests, non-exclusive.

repositories from two software companies (266 and 138)
404
lines of code by the owners' automated reports (5,011,228 and 4,429,008)
more than9.4M
merged pull requests, with more than 206,000 commits
nearly119,000
repositories delivered with their support tickets, engineering chat, work tracking and error data
30

What is inside

  • second software company: 266 repositories, 76% JavaScript, 20% TypeScript and 2% Python by lines, earliest commits from 2016 to 2026; 92 run CI and 7,339 test files are committed.
  • 30 of those repositories, 4,229,247 lines, come with the support desk, engineering chat channels, work-tracking records and error-monitoring data; links between tickets and commits have not been checked yet.
  • Sales-software company: 138 repositories after the owner's fixes, with C# the main language in 87 of them (47.7% of all lines are C#); the main application has 2,888,550 lines over about 12 years to October 2026, 75,186 commits, 38,849 merged pull requests and 193 contributors.
  • 45 of the 138 changed in the last 90 days, 55 have 100 or more merged pull requests, and 13 carry a training-task range; the second software company adds 14 task-grade repositories.
  • AI assistance is disclosed per repository: AI-assistant files are committed in 28 of the second company's repositories and 23 of the sales-software company's, and in 2 of the latter an AI coding agent is a commit author.
  • Keys, passwords and personal emails found in code are removed before delivery.
  • Figures come from automated reports the owners ran on each repository, which share metrics and no code; generated and vendored code is not counted, and nothing has been measured by Rightmark yet.

Best for

  • Training and evaluating coding agents on multi-year company codebases in two stacks (C#; JavaScript and TypeScript)
  • Building software-engineering tasks from real commits and merged pull requests (27 task-grade repositories flagged)
  • Issue-to-fix and ticket-to-commit research on the 30 repositories delivered with their tickets
  • Code review and pull request models

Before delivery

  • Each owner confirms its rights to the delivered repositories; the sales-software code is offered only once its owner holds IP assignments from its engineers and contractors.
  • Repositories that were ever public, open-source and third-party code, second copies, client material and generated or vendored code are left out.
  • Keys, passwords and personal emails found in code are rotated or removed before delivery.
  • AI-assistant files and AI-agent commits are disclosed per repository, so a buyer can leave them out.
  • Ticket-to-commit links on the 30 linked repositories are measured and the share stated before delivery.
  • Each owner signs its own license; the buyer gets one card and one order. Delivery starts once two members have cleared title.

Left out

  • Open-source and third-party repositories, second copies, client material and keys, generated or vendored code.
  • Two second company's repositories whose files carry copyleft license headers, until a rights review clears them.
  • Any repository confirmed as ever public.
  • Personal data in the linked tickets and chat, masked per the privacy review.

Lot BND-3-RECORDS

How two B2B software companies work: more than ten years each of mail, chat, wikis, tickets and CRM

One license covers two companies' records: a sales-software company (about 300 to 350 mailboxes over the years, team chat from 2015 to October 2026, Jira, Notion, Intercom and HubSpot) and a second software company of a few hundred staff (more than 3.5 TB of mail and files, more than 72,000 support tickets, more than 590,000 CRM records). Exclusive for 24 months or non-exclusive.

of records at each company: one sales-software company and one other B2B software company
more than10 years
of mail, documents and chat at the second software company, with about 4.09 million chat messages in the last 13 months
more than3.5 TB
CRM records and more than 72,000 support tickets at the second software company
more than590,000
mailboxes over the years at the sales-software company, with team chat from 2015 to October 2026 (about 2.18 million messages by the owner's count)
about300 to 350
at the sales-software company: a mail and documents suite, team chat, a wiki, work tracking, a support desk, a CRM and a call recorder
7 tools

What is inside

  • Sales-software company: email and documents, team chat, a wiki, work tracking, customer support and CRM.
  • About 300 people have worked there, 50 to 120 at a time, about 75 today; most staff work outside the US.
  • second software company, more than ten years: mail and documents (3.53 TB), team chat (about 60 GB, from 2015), 72,566 support tickets, 591,074 CRM records and about 100,000 calendar records.
  • A few hundred staff; the owner estimates that about 70% of the records come from staff in India and 30% from the US. Its records are in English.
  • Recorded calls (up to 15,900 hours at the sales-software company) and both companies' code can be added; each is priced on its own card.
  • Delivery only after a rights and privacy review; samples come only with each owner's agreement, as metadata or masked output.

Best for

  • Enterprise environments and agent training on real cross-tool work: mail, chat, wiki, tickets and CRM at two companies
  • Testing agents across two company shapes and regions
  • Long-horizon customer support models on more than ten years of tickets and support chat
  • Sales and customer-success workflows inside a sales-software company's own CRM and support desk

Before delivery

  • Each company's records pass a rights and privacy review before delivery; personal data is masked per that review.
  • Each owner states its exclusions in writing; always out: health data, under-18s, payment-card and bank-account numbers.
  • One exclusive per company: if a company's records are licensed exclusively elsewhere before this bundle is signed, that company leaves the bundle.
  • Code and calls already licensed non-exclusively are listed as carve-outs before signing.
  • Volumes per tool, and the year each tool's records start, are measured and stated before contract.
  • Each owner signs its own license; the buyer gets one card and one order. Delivery starts once two members have cleared title.

Left out

  • Always out: health data, data about under-18s, payment-card and bank-account numbers.
  • Code and recorded calls, priced on their own cards; either can be added.
  • Private chat channels at the sales-software company, unless its owner adds them.
  • Any records a tool vendor claims rights over.

How to quote

Name the lot and the license

Quote at the ask for an exclusive license, a non-exclusive license or an exclusive window, propose your own terms, or pass.

Samples under NDA

Samples and detailed source lists come through Rightmark's buyer portal once you accept its NDA, and only with the owner's agreement.

Rights and privacy, lot by lot

Each owner's rights are reviewed before delivery, with a privacy review for every lot, and the counts are measured and listed in a manifest before signing.

Write to us

Write to hello@rightmark.ai and name the lot, or send a request.

For data buyers

Need data that is not listed?

Tell us what you need and we reply with scope and timing.

Request data

For data owners

Own data like this? List it with Rightmark.

Start with an indicative appraisal: no files or system access needed. It is free, and nothing is licensed until you sign.

Appraise your data

Or start from company records or source code. Know an owner? Become a Scout.