Monetize your old datasets. The time is now.
Companies and projects fail or get outdated. The work that went into them is still worth a lot of money.
We help you recover costs by selling your datasets and IP to AI foundational models. We anonymize the data and handle legal. First call to payout in four weeks or less.
What we collect
Startup data and legacy codebases.
Startup data.
From companies that operated for two or more years and have wound down or are winding down. The messy reality of how a real business ran is what public training data cannot replace: emails, decisions, transactions, design files, and the back-office systems that ran it.
View the criteriaHide the criteria
Headcount
10+ at peak. More data, more value.
History
Operated for two or more years. Enough timeline to be interesting.
Type
Real software, tech, or digitally-operated business. Not pure services.
Authority
Founder, investor, or estate signatory authorized to transfer.
Obligations
No ongoing customer obligations blocking disclosure.
View the full listHide the list
Communication
Email, chat, Slack, Teams, Discord, voicemail.
Project tracking
Jira, Linear, Asana, Monday, ClickUp, Trello.
Documents
Drive, SharePoint, Dropbox, Notion, Confluence.
CRM and sales
Salesforce, HubSpot, Pipedrive, deal notes, lost-deal reasons.
Customer support
Zendesk, Intercom, Freshdesk, ticket transcripts.
Meeting data
Zoom, Gong, Granola, Otter recordings and transcripts.
Financial
QuickBooks, NetSuite, Xero, ERP exports, P&Ls, AR and AP.
HR and recruiting
Greenhouse, Lever, performance reviews, calibration notes, comp data.
Design and product
Figma files, PRDs, design reviews, spec docs.
Analytics
Looker, Tableau, Mixpanel, Amplitude exports.
Internal wiki
Runbooks, onboarding docs, internal training material.
Strategy and board
Board decks, OKRs, strategy memos, all-hands recordings.
Legal and contracts
Contract templates, redlines, vendor agreements.
Legacy codebases.
The codebase no one wants to maintain is the one buyers will pay most for. Frontier teams need agents fluent in the stacks that still run banks and infrastructure. Public training corpora missed almost all of it. Production code, repository history, comments, and issue trackers all add value.
View the criteriaHide the criteria
Authenticity
Real production code, not toy projects or tutorials.
Size
50K lines or more. Diminishing returns past 1M.
Docs
Comments, internal documentation, README files where they exist.
Repo history
Version control where available, including git, SVN, CVS, or Perforce dumps.
Issue tracker
Jira, Bugzilla, or Mantis dumps are a strong multiplier.
Build
Test suites, build scripts, deployment configs.
IP chain
Clear ownership. Sole owner, ex-employer with rights, estate, or university policy.
View the language listHide the list
COBOL
Banking and insurance core systems.
FORTRAN
Scientific computing, oil and gas, defense, weather modeling.
RPG / IBM i
Mid-market ERPs, manufacturing, distribution.
Ada
Defense, aerospace, rail signaling.
Pascal / Delphi
Industrial software, point of sale, Central European enterprise.
PL/I
IBM mainframe, financial services.
PowerBuilder
Client-server enterprise apps from the 90s and 2000s.
Visual Basic / VBA
Internal tools, Access apps, Excel macros at scale.
ColdFusion
Older enterprise web applications.
Perl
Sysadmin, bioinformatics, legacy web.
Smalltalk
Banking and trading systems.
ABAP
SAP customizations.
Legacy Java EE
Pre-Spring-Boot enterprise Java, J2EE, older application servers.
Why now
Public training data is exhausted. Proprietary data is not.
Frontier buyers have ingested most of the public web. The next generation of agents needs workflows and code that were never published.
Central Europe sits on three decades of work no one in San Francisco knows how to reach. Banks, telcos, manufacturers, and the long tail of startups that did not make it.
The window is short. Buyers are bidding now. Legal, anonymization, and placement take weeks. Most archives in basements will still be there in two years; most buyers will not.
Process
First call to payout in four weeks or less.
Week 1
Initial call. NDA signed. Data assessment.
Week 2
Price offering. Legal contracting.
Week 3
Anonymization. Transfer to chain of custody.
Week 4
Data pack transferred. Money wired.
Why us
We bridge the CEE startup ecosystem with foundational models.
01
CEE ecosystem builders
We have been in the Central European tech and startup ecosystem for more than fifteen years. Matej Ftáčnik, one of our founding partners, co-founded The Spot, the first founder community of its kind in Slovakia, and built Vacuumlabs into one of the leading software engineering practices in the region. The lawyers, founders, and engineers we call are people we have known for a decade or longer.
02
US foundational model relationships
Our partnership with an American-based intelligence company gives us a direct line to Anthropic, OpenAI, and Apple. Our deep relationship with them lets us move fast and operate on mutual trust.
03
Fair game
We are founders too. We know how much goes into building companies and how little comes back when it doesn't work out. More money back to CEE founders means more resources in the ecosystem. That is the bigger picture we are working toward.
04
Operational discipline
We move fast. The first conversation is thirty minutes under NDA. A letter of intent lands inside the first week. The wire clears by week four. We do not run open-ended diligence, we do not put names on lists, and we do not stretch a process that should take days into weeks of theater. The work is yours, the wire is yours, and the timeline is short.
What we handle
We make sure everything is ready for a fair, clean & correct sale.
Anonymization
PII, customer data, and trade secrets are removed by a vetted partner before the data leaves the chain of custody. You never ship raw archives to the end buyer.
IP and chain of title
We draft reps and warranties. Sole owner, ex-employer with rights, estate signatory, university policy. We work through the chain so the seller does not.
Compliance
GDPR and sectoral rules are our responsibility. The regulatory burden does not transfer to the seller.
Let's monetize your old datasets. The time is now.
Tell us what you have in terms of data, or recommend someone who might benefit from this. We make sure to respond ASAP.
Expect a reply within 2 working days.