What we store, and what we do not

A paper you upload stays private to your own workspace. It is not added to the public library, and it is not shown to other readers, unless its licence permits republication and you publish it yourself.

The public library contains open-access and CC-licensed work only.

If you create an account we store your username, display name, bio, and the answers you gave during onboarding.

Who can read your onboarding answers. A Pepiros administrator can read them, including the free-text ones: what an AI summary got wrong for you, how you check a claim today, what you actually work on, and what would make you use this weekly. They are read to work out which problems are real and what to build next. They are not sold, not shared with third parties, and not published.

Email contact is opt-in. The checkbox on the last onboarding step is off unless you tick it, and we do not email you about your reading unless you do. To withdraw it, untick the box (the wizard is resumable at any time) or email us and we will clear it, along with any free-text answers you would rather we did not keep.

Everything else you answered is segmentation (role, field, intent, experience, referral, country) and decides what your home page surfaces first.

The guest demo account stores nothing. Everything it shows is generated, and nothing you do inside it is written anywhere.

A signed-in session lasts 7 days and does not silently renew. Signing out, or signing out everywhere in settings, revokes it immediately rather than leaving it valid until it expires on its own.

Pepiros calls Semantic Scholar to populate the related-papers rail and OpenAlex to expand a citation graph. Both receive the identifier of the paper being looked up. Neither receives your account details.

Grounded chat sends the text of your question and the relevant source excerpts to a language-model provider. It does not send your identity.

There is no third-party analytics or advertising script on this site. The reach figures on your own dashboard are counted server-side and are visible only to you.

Privacy questions go to the same place as everything else: open an issue on the repository. For anything sensitive, follow the disclosure process on the security page instead of a normal issue.