Privacy statement

We do not identify you. To count the free documents you can open, your browser keeps a cookie listing the documents you viewed (at most ten, signed so it cannot be altered, kept 30 days). We record what is done on the site — searches, clicks on search results, document views — without any identifier, so these records say what was searched, not who searched it. Recent searches and documents you opened are remembered in your own browser's storage and never sent to us.

We store:

• Your email address, a display name (if you leave it empty, the part of your email before the @), and your password in hashed form.
• The questions you ask the AI and its answers, and your conversations. An answer can be reused for a very similar question from another user within 24 hours; your identity is never shown with it.
• Documents you save and your reading progress, and your citation-style preference.
• Annotations. Public annotations and references show your display name to other readers; private notes are visible only to you.
• Answers you choose to share. They appear on the public research feed with the question and answer but without your name, and you can remove them.
• Your searches, clicks on search results, document views and downloads, linked to your account. We use this to improve search and to see which documents are useful.
• How much you use the AI (requests and tokens per day), to apply usage limits.
• PDFs you upload. They become documents in the public library; who uploaded them appears only in our server logs (31 days).

Payments are handled by Mollie B.V. (the Netherlands). Mollie receives your name and email address to create a customer record, and your payment details go to Mollie directly — we never see or store card or bank account numbers. We store Mollie's reference numbers, the status, amount and dates of your subscription and payments.

If you connect an AI assistant to the library through MCP, we store your API keys and access tokens in hashed form, the apps you authorised, and every request the assistant makes: which tool it used, the request it sent (for example the search query), whether it succeeded and how long it took. When an app registers itself for MCP access, we keep the IP address it registered from, to stop abuse of the registration endpoint.

The newsletter is separate from an account. We store your email address, whether you confirmed it (double opt-in) and where on the site you signed up. Every issue has a one-click unsubscribe link. After you unsubscribe we keep your address marked as unsubscribed so we do not email you again; ask us if you want it removed entirely.

• Account emails: confirming your address and resetting your password.
• The newsletter and its confirmation email, if you signed up.
• When you create an account, the site owner receives a notification with your email address and display name.

Your IP address is used only briefly, in memory, to limit how many requests one visitor can send; it is not stored with your activity. Our server logs record requests without IP addresses and are kept for 31 days. Our content delivery network (Cloudflare) processes your IP address to deliver the site and protect it from attacks.

• Sign-in cookies (names starting with sb-), set when you log in, to keep you signed in.
• vault_doc_views, for visitors without an account: the documents opened, to count free reads (30 days).
• Browser storage on your own device for preferences (citation style, sidebar width, highlight colour), your recent searches and documents, and a cached copy of the glossary. This is never sent to us.

We use no advertising, tracking or analytics cookies, and no third-party scripts that follow you across sites.

We use these services to run the site:

• Hetzner (Nuremberg, Germany): the servers and database the site runs on.
• Cloudflare: content delivery and protection for all traffic, storage of uploaded documents, and sending app emails.
• Amazon Web Services (SES): sending account emails.
• Mollie (the Netherlands): payments.
• OpenRouter, and the AI model providers it routes to (such as Google): your AI questions and the conversation they belong to, to generate answers.
• Jina AI: search queries and AI questions, to find and rank relevant documents and to search the web for an answer.
• Tavily: web search queries, when Jina’s web search is unavailable.
• Steam, IGDB/Twitch and PandaScore: only the game names the AI picks out of your question, to add live game data. No personal data.

Some of these providers are located outside the European Economic Area, mainly in the United States. Where that is the case, the transfer relies on the safeguards those providers offer for it, such as the EU–US Data Privacy Framework or standard contractual clauses. We do not sell your data or share it for advertising.

• To provide the service you signed up for — your account, AI answers, saved documents, subscription (performance of a contract).
• To keep the site secure, prevent abuse and improve search and the library (our legitimate interest).
• To send the newsletter (your consent, which you can withdraw at any time).
• To meet legal obligations, for example for payment records.

• Account data: as long as your account exists. Deleting your account removes your questions and answers, conversations, saved documents, annotations, shared answers, usage records, MCP keys and connected apps, and your subscription and payment records with us.
• Activity records (searches, clicks, views, downloads, MCP requests): at most 400 days. When you delete your account, the ones linked to you are kept only without any link to you.
• Server logs: 31 days.
• Newsletter: until you unsubscribe or ask us to remove your address.
• Payment records and invoices are kept for 7 years by Mollie, which handles our payments and invoicing, and in our bookkeeping, as Dutch tax law requires — also after you delete your account.

An email address and a password are needed to create an account, and a subscription needs a payment through Mollie; without them we cannot provide those services. Everything else — a display name, saving documents, annotations, the newsletter — is up to you.

We make no decisions about you based solely on automated processing, and we do not build profiles that have legal or similarly significant effects on you. The sources an AI answer draws on can be ordered partly by the topics of your recent questions; that only changes which documents the answer looks at first.

You can ask to see the data we hold about you, have it corrected, deleted or handed over, and object to or restrict how we use it. You can delete your account yourself on your account page; for anything else, email [email protected]. If you think we handle your data wrongly, you can complain to the Dutch Data Protection Authority (Autoriteit Persoonsgegevens).

When the site starts storing or sharing something new, we update this page and its date.