Privacy & copyright
Plain-language summary of how the service handles your data. This describes intended behavior of the product as built; it is not legal advice.
What we store
Your account (name, email, a hashed password), the texts you upload and the text extracted from them, and everything you create: highlights, notes, bookmarks, reading status, understanding ratings, and any corrections you make to automated annotations. Preferences such as your expertise level are stored to personalize the roadmap.
Your data is isolated to you
Every piece of your content is scoped to your account and never visible to other users. Requests for a resource you don’t own return “not found” rather than revealing it exists.
Copyright — you bring your own texts
The service never bypasses paywalls or sources copyrighted material on your behalf. For works it can’t obtain, it stores a proper citation and a pointer toward legitimate acquisition, and lets you upload your own legally obtained copy. Roadmap and analysis still work for works not directly inspected, marking them accordingly.
Automated classification
Bibliographic facts come only from real lookups (Crossref, OpenAlex, Open Library), never generated by a model. When a classification model is used to classify how references relate to a text, calls go to the provider’s API tier under its data-use terms. Your uploaded content is never used to train models without your explicit, separate opt-in — the default is opt-out.
Deletion & export
Deleting your account removes your uploaded files, extracted text, annotations, and derived data — not just the account record. You can export your notes, roadmap, and bibliography.
A research aid, not settled scholarship
Every system-generated claim carries a confidence and its provenance, explaining the basis for the inference rather than presenting it as settled fact. Verify against the primary text before relying on anything; you can correct, dispute, or hide any automated annotation.