We provide frontier labs and AI companies with data from established European firms. It supports training, evaluation, and product development. Anonymized and unavailable on the open web.
Shared drives, docs, invoices, forms, CRM records. Historical data from established firms.
- • Correspondence
- • Invoices
- • Contracts
- • Ledgers
- • Forms & CRM
Data produced every day by workers. Fresh, operational, real-world patterns.
- • Daily transactions
- • Process logs
- • Communication
- • Task tracking
- • Status reports
Archives
Coherent document estates from established sources (correspondence, contracts, invoices, ledgers) with documented chain of custody.
Structured extracts
Entity-resolved, anonymized extractions from archive material, delivered against your schema and validated per field.
Evaluation sets
Held-out, never-published document tasks for benchmarking extraction, reasoning, and agent workflows on genuinely unseen data.
Operational records enter through cleared channels. Nothing is scraped. Every corpus is authorized and anonymized at the source.
Insolvency estates
Document estates from wound-down European companies, which administrators release only under a written mandate. Decades of correspondence, contracts and ledgers, no longer behind a firewall.
Workflow capture
How expert work actually gets done, recorded as step-by-step trajectories while the work is happening, with consent. Operational knowledge the open web never held.
Personal documents
A contributor app, not yet released, lets individuals hand over their own paperwork, anonymized on-device before anything ever leaves the phone.
Bespoke corpora
Need something the channels don't already hold? We hunt and assemble professional datasets to your spec: healthcare, legal, finance and beyond.
For AI teams
Corpora are licensed for pretraining, fine-tuning, and evaluation. Each delivery includes a manifest, anonymization report, and license terms.
- RightsWritten chain of title from the originating estate to your license. No inferred permissions.
- PrivacyAnonymization verified per batch and documented in the delivery report.
- ExclusivityCorpus-level exclusivity windows available per license.
- DeliveryS3 transfer or physical handover. Checksummed manifests on every corpus.
Your archive is an asset
We work with companies and institutional sources to license dormant document estates under written authorization, with revenue share on every license and zero operational lift on your side.
- AuthorizationWritten mandate before a single file moves. You define the scope.
- ConfidentialityAnonymization before any third party sees content. Named entities never leave the pipeline.
- RevenueShare on every license of your corpus, reported per transaction.
- LiftWe handle capture, clearance, staging, and sale end to end.
- 1.0AcquireSourced from company archives and institutional sources. Never without written authorization.
- 2.0ClearRights review per corpus: ownership, third-party material, statutory limits. Documented, not assumed.
- 3.0AnonymizeNamed entities, personal data, and commercial identifiers removed or pseudonymized, verified per batch.
- 4.0DeliverStaged corpora with manifests, checksums, and a license your counsel can actually sign.
Get in touch
Whether you need data, hold an archive, or can introduce us to one, tell us what you're working with and we'll respond within 48 hours. Introductions that lead to a corpus or a partnership earn a referral fee.
Contact →