Skip to content

Model catalogue · Mistral AI

Mistral OCR

Mistral OCR reads documents - scanned pleadings, invoices, tables, handwritten notes - and returns structured text including layout; without this step no model can work on a file that exists only as an image. It is procured via Google Vertex AI, and there the chain is closed: the platform operator's signed confidentiality agreement under section 203 of the German Criminal Code is service-bound and expressly covers third-party models as well, with processing during the request committed to the EU and to the Netherlands. That weighs particularly heavily here, because an OCR pipeline passes through complete files rather than extracts.

Through us once the provider clears it

The confidentiality agreement under section 203 of the German Criminal Code is signed. We set the route up without storage at the provider (zero retention) once the provider has committed to it - at Microsoft through the exception from abuse monitoring, which usually takes months. Until then, storage and review by the provider are not ruled out.

What it is good for

  • Reading scans, receipts and tables with layout preserved - the preliminary step for all file work
  • Processing committed to the EU and to the Netherlands, European model vendor
  • Covered by the platform operator's service-bound section 203 agreement

Where it stops

  • An OCR pipeline sends complete files through the service, not extracts - prior redaction is practically impossible here, which makes the contractual position all the more important
  • The platform operator is a US group - CLOUD Act exposure remains despite the European model vendor
  • The reach of the logging exception for third-party models cannot be evidenced from publicly available documents (as of 2026-08-17)
  • Recognition errors on handwriting and poor scans remain - figures and deadlines from an OCR run have to be read back
  • The platform's global endpoint gives no residency guarantee - the same models can be called there as well. Binding to the EU is therefore a configuration requirement: the global endpoint is blocked by organisational policy, otherwise the EU commitment is a settings question rather than a property.

Status through us

The same model, procured differently, is a different legal position. The status above refers to procurement through us; processing takes place in a third country.

Which route carries it, with which instrument and under which condition, we set out in full to the firm before it commissions us - section 203(4) requires that the chain can be checked, not that it is published. The full matrix, with the sources behind every route, can be requested in the German catalogue after confirming your e-mail address, and we go through it in the conversation.

Is Mistral OCR the right fit?

We go through your tasks and the route that carries them - and say plainly where it does not.

Arrange a conversation