Privacy Policy

• Account information. You sign in with Google. We receive your name, email address, profile picture and Google account identifier from Google and use them to create and identify your account. You can add a phone number in your profile.
• Your content. Documents you write, PDFs and other files you upload, papers you add to your library, AI chat messages and instructions, and the version history of your documents.
• Billing information. If you subscribe, payment is handled by Stripe. We receive your subscription status and Stripe customer and subscription identifiers; we never see or store your full card number.
• Usage and technical data. A record of each AI request you make, without its content (the feature, the model used, its size, cost and response time), which we use to apply plan limits and keep the service reliable, and server logs that include your IP address and the requests you make, which we use to run and secure the service and keep for 30 days.
• Product analytics. How Papernex is used, as described under "Analytics" below.
• Bug reports. When you report a bug in the app, what you write, a screenshot if you attach one and, unless you turn them off, technical details: the page you were on, the app version, your browser, screen size, language and time zone, your workspace and open document, and your last errors and failed requests (which part of Papernex was called and how it answered, never what was sent, passwords or sign-in tokens). You see these details before you send the report. Only Papernex administrators read bug reports, to fix the problems they describe.

We use your information to provide and improve Papernex: to store and sync your documents, run AI features on your content, enforce plan limits, process payments, respond to support requests, and detect abuse or security problems. We do not sell your personal information and we do not use your content to train AI models.

When you use AI features (chat, writing commands, autocomplete, reviews, paper insights, research workflows, slides and data results, chat titles) or upload a PDF (to read its title, authors and other details), the text that feature needs from your documents, chats and library is sent to the AI service providers that run our models, to produce the response, and to Microsoft Azure OpenAI Service to index your papers for search. These providers include Microsoft Azure OpenAI Service and Cloudflare Workers AI; other providers we configure may process requests under equivalent terms. We use providers that process this data to produce the response and do not use it to train their models; they may keep requests for a limited period to detect abuse, and requests may be processed in data centers outside your country. We never use your content to train AI models.

Some features look things up in public scholarly services. When you search for papers, cite by DOI, ISBN, arXiv or PubMed ID, add a PDF, or run a review, search terms and identifiers (for example a paper's title, authors or DOI) are sent to services such as OpenAlex, Crossref, arXiv, doi.org, PubMed (NCBI), Open Library and Google Books, and when you cite a web page we fetch that page. The AI assistant and reviews may write these search terms from your question or from claims in your document, so they can contain a few of its words. Your documents themselves are never sent to these services.

When you search the web in Papernex (the citation picker's Web tab, Find papers, or AI chat with web search turned on), the search terms go from our servers through our own SearXNG search service to third-party search engines and scholarly indexes, such as Google Scholar, arXiv, PubMed, Semantic Scholar, Crossref, OpenAlex, Wikipedia, Mojeek, Yahoo, Yep, DuckDuckGo and Bing. In chat, the assistant writes the search terms from your question, so they may contain words from it or from your document. The search engines receive only the search terms: never your name, email address, account or other user identifiers, IP address or documents. Web search in chat is off until you turn it on for a chat. We don't store search terms with your account; results are cached for a few hours under a one-way hash of the search.

We share information only with the providers that run Papernex for us:

• Microsoft Azure hosts the app, database and file storage in the United States, sends our notification emails, and runs AI models and search indexing (Azure OpenAI Service).
• Cloudflare runs some of the AI models behind Papernex's AI features (Cloudflare Workers AI).
• Other AI service providers we configure may run AI models for these features, under equivalent terms, as described under "AI Processing".
• Google provides sign-in. Google's sign-in button and the app's fonts are loaded from Google, which receives your IP address when they load, and are covered by Google's privacy policy.
• Stripe processes payments and subscriptions and is covered by Stripe's privacy policy.

We may also disclose information when required by law or to protect the rights, property or safety of Papernex, our users or others.

We measure how Papernex is used with our own, first-party analytics. We do not use third-party analytics trackers, advertising pixels or session recording.

• On papernex.com we count page views without cookies: the page's path, the referring website, campaign tags (utm_source, utm_medium, utm_campaign) and the device type, browser and operating system family. We do not store your IP address or any identifier, so visits cannot be linked to you or to each other.
• In the app we record usage events tied to your account: pages you visit, features you use (for example citations inserted, exports, AI actions and imports), session length, page performance measurements and error reports. We use them only to understand and improve Papernex. They never include the content of your documents, chats or sources.
• If your browser sends a Do Not Track or Global Privacy Control signal, neither the website nor the app sends any analytics. The service still keeps the records it needs to run your account, such as when you last signed in or were active.

The app uses a cookie to keep you signed in, and a random session number kept only for the open browser tab to group its analytics events. It also keeps a few things in your browser's local storage: your name, email address and photo (so the app opens quickly), display preferences such as the theme and panel sizes, revision plans you saved for a document, links to the share links you created, and, if you connect Zotero, your Zotero API key (only for the browser session unless you choose to remember it on that device). The website remembers your light or dark theme choice in your browser. We do not use advertising cookies or third-party analytics trackers.

We keep your account and content while your account is active. When you delete a document, it is removed from your workspace right away and erased for good when its workspace is deleted. When you delete a PDF or other file, it is erased right away.

When you delete a workspace, everything in it is deactivated at once and permanently erased after 30 days; until then you can ask us to restore it.

You can delete your account yourself in the app (Settings → My account → Delete account). This immediately:

• deletes your profile (name, email address, phone number and photo), your Google sign-in record and sessions, your AI chats, personal saved prompts, notes and highlights on papers, reading status, AI-use declarations, notifications and their settings, emails waiting to be sent to you, and the record of the days you were active;
• cancels your subscription by deleting your customer record at Stripe, without a refund for the current period, and stops share links you created;
• deletes the workspaces only you own, for their other members too (they are told, and everything in them is erased permanently after 30 days), and removes you from the other workspaces you belong to;
• unlinks your product analytics events from you, so they can no longer be tied to your account.

What you contributed to workspaces owned by others (documents, papers, comments and shared prompts) stays there, shown as written by "Deleted user". Your account is kept only as an anonymous placeholder so that this content and our records of AI requests (which never include their content) still work; while we keep the deletion record below, it can be connected to you. Signing in again with the same Google account creates a new, empty account. You can also ask us to export or delete your data by emailing info@papernex.com from the address you sign in with; we complete requests within 30 days.

Bug reports you sent us (what you wrote, any screenshot and the technical details included with them) are kept so we can fix the problems they describe. They stay linked to your deletion record below, so while we keep it they can be connected to you.

Deletion record. When an account is deleted we keep a record of the deletion for 2 years, then delete it automatically. It holds the account's email address and name, how it signed in (for example Google), when it was created and deleted, who deleted it (you, or us at your request), the reason and comment you gave, your plan and subscription status at the time, your Stripe customer ID and whether it was deleted at Stripe, how many workspaces, documents, papers and chats were deleted or left behind, how many other members lost access and which workspaces were deleted (by their internal ID), the network part of your IP address (the first three of its four numbers, or the first 48 bits of an IPv6 address; never the full address) and your browser, operating system and device type. We keep it to answer support requests about a deleted account, to prevent fraud and abuse, to handle billing disputes and chargebacks, and to meet our legal obligations. Only Papernex administrators can see it.

Deleted data can remain in our database backups for up to 14 days, and deleted files in our file storage's recovery copies for up to 30 days, before they are erased. We also keep records we must keep by law (for example, payment records held by Stripe).

Data is encrypted in transit (HTTPS) and at rest in Microsoft Azure. Stored files are private and served through time-limited signed links, and access to production systems is limited to the people who operate the service. No system is perfectly secure, but we work to protect your information.

Depending on where you live, you may have the right to access, correct, export or delete your personal information, or to object to or restrict how we use it. Email info@papernex.com to make a request. You may also complain to your local data protection authority.

Papernex is not intended for children under 13, and we do not knowingly collect personal information from them.

Papernex links to external sites such as publisher pages and DOI links. Their privacy practices are their own; please review their policies.

We may update this policy. We will post the new version here with a new effective date and, for significant changes, let you know in the app or by email.

Questions about this policy or your data: info@papernex.com.