Trust
Built to pass your legal and data review.
Every dataset is prepared the way a legal team, a data protection officer and an ML lead would want to receive it.
01
Rights chain
Written permission from each contributor, kept with the dataset.
02
De-identification
Client, patient and personal data removed. Technical content kept intact.
03
Data protection
Medical data released only after a data protection review, under NDA.
04
Provenance
Origin, language and source type stated in every manifest.
05
Verified files
Opened and checked in native software before delivery, with checksums.
06
Clear terms
Non-exclusive by default, exclusive when agreed. One contract, one counterparty.
What we never do
- Scrape data from the web
- Offer files without written permission from the owner
- Share patient data that has not been de-identified
- Name a data owner or a buyer without their consent
What every buyer receives
- A manifest listing every file, with checksums
- The origin and language of each collection
- The licence terms and permitted use in writing
- A single counterparty for contracts and payments