SECURITY
Your documents are the most sensitive thing you will give us.
We treat them that way. This page sets out how documents are handled, in plain terms.
A private cloud, not a shared API call
Extraction runs on a small, purpose-built model in a private cloud environment dedicated to your data, rather than sending documents through a shared third-party API. Your documents and the data extracted from them are never used to train any model, for your account or anyone else's.
WHERE YOUR DOCUMENT
The same document, the same answer, every time
Extraction is deterministic and rules-based. There is no generative guesswork in the extraction layer, which means the answer you get today is the answer you would have got last month on the same document, and the one you will get next year.
WHY DETERMINISTIC MATTERS HERE
Access is scoped to your account
Documents and the database built from them are held separately for each client and are not shared or pulled across clients. Access is controlled per user, and full source document access is included for anyone entitled to see it. Retention and deletion terms are agreed as part of onboarding and can include deleting source documents once the database is built, when that is what the client wants.
ACCESS AND RETENTION
