Where it belongs
Mathpix is configured at the platform level and exposed to premium workspaces from the asset page. That keeps OCR attached to source material instead of mixing it directly into items or documents.
The extracted result is stored on the asset so it can be reviewed, copied, and reused later while building educational content.
What gets stored
Didaxi stores the OCR result it needs on its own side, including plain text and LaTeX. This avoids relying on the provider as durable storage.
The extraction result should be treated as a draft. Review it before turning it into items or public material.
Privacy defaults
The image is sent to Mathpix from the server, not from the browser, and the request is made with retention-minimizing metadata by default.
That keeps private creator material under tighter control while still allowing extraction.