1.What this addendum covers
Document Reading lets AI clients you connect to your account read your reMarkable documents and search their contents. To make that possible, we keep a copy of your document content on our servers. This is a change from our default posture — for everything outside this feature, we handle document contents in transit only and do not retain them (see our Privacy Policy).
The feature is off by default. Nothing described in this addendum happens until you explicitly enable Document Reading from your Document Reading settings, and it stops — with the deletion described in §6 — when you disable it.
2.What we store
While Document Reading is enabled, we store:
- Your document content. The pages of documents synced from your reMarkable cloud account, identified by document name, page number, and the page's content hash.
- Text we extract from it. Markdown text recovered from each page — typed text read directly from the page data, and handwriting transcribed as described in §3 — together with a search index built from that text.
This data is stored in our Postgres database, hosted with PlanetScale, Inc. (already listed as a service provider in our Privacy Policy), and is accessed only by our Worker on behalf of your account.
3.Handwriting transcription
Pages that contain handwriting are rendered to an image and sent to OpenAI, our AI model provider acting as a sub-processor, which returns a text transcription. We store the returned text alongside the page it came from. The provider processes page images solely to produce these transcriptions for us.
Automated transcription is imperfect. Transcribed text may contain errors, and we make no guarantee of its accuracy or completeness.
4.One-way sync
Document Reading only reads from the reMarkable cloud. It never writes to your tablet or your reMarkable cloud account. The only thing the Service ever writes to your tablet is a document you push yourself through an AI client you have connected — exactly as before this feature existed.
5.Retention
We keep the stored content and extracted text for as long as both of the following are true: your plan is a paid plan, and Document Reading is enabled on your account. When either stops being true, the deletion described in §6 applies.
6.Deletion
We delete all document content and extracted text stored under this addendum immediately when any of the following happens:
- you disable Document Reading from your Document Reading settings;
- your plan changes to the free plan (including cancellation of a paid subscription); or
- you delete your account.
Deletion removes the stored pages, the extracted text, and the search index built from them. Your documents on your tablet and in your reMarkable cloud account are not affected — we only delete our copy. If you enable the feature again later, content is synced fresh from your reMarkable cloud.
7.Relationship to the Terms and Privacy Policy
This addendum supplements our Terms of Service and Privacy Policy; both continue to apply in full. Where this addendum speaks specifically to content stored for Document Reading, this addendum controls. If we make material changes to this addendum, we will post the revised version here with a new version identifier and ask for your consent again before the changes apply to you.
8.Contact
Questions about this addendum: Joshi LLC (doing business as Remarkdown LLC) — support@remarkdown.org