Privacy policy

Your account. Your name, email address, a hash of your password if you set one (never the password itself), your settings, and your sign-in sessions (when they started and how you signed in). If you pay, your plan and invoices; card details stay with Stripe.

Your researcher profile. After you first sign in we ask for your name, your institution, university or company (or “N/A, independent / not affiliated”) and your role (student, postdoc, faculty, librarian and so on), and you may add your field and your ORCID iD. We use it to know who an account belongs to, to help you when you contact support, and to count, in aggregate, which kinds of institutions and roles use Citation Monster. Your colleagues in a workspace see your name and affiliation; your role, field and ORCID iD are seen only by you and our support staff. It is never used to train AI models. While you type your institution, the text is sent to OpenAlex to suggest matches. You can change it in Settings. A workspace export you request includes your whole profile (and your colleagues' names and affiliations), and the workspace's owners and admins can download it. Your profile is deleted with your account.

Your uploads and private work. The PDFs you upload, the text and figures we read from them, the references we extract and check, and everything you make in a workspace: highlights, notes, chats with your questions and the answers, research projects and runs, knowledge bases, folders, saved prompts, add-ons, custom assistants and sources (connectors), and exports. This is private to your workspace. Passwords and keys for sources you connect are encrypted with a key that belongs to your workspace.

Publicly available scholarly material we download, process and store. To find, check and link references we download metadata (titles, authors, DOIs, abstracts, citation counts, open-access links) from public sources such as OpenAlex, Crossref, Unpaywall and Europe PMC, and the open-access PDFs of cited works. We store this material, read and process it (for example its text and its own references), and keep one shared copy that every user's papers can link to. Because it is public it is not treated as your data, and it is not deleted when you delete your account. What stays private is your connection to it: which of your papers cite a work, what your paper says about it and the results of your checks.

Usage and security records. Server logs (with paper text, prompts, answers, emails and keys removed before they are written), a security audit log of sensitive actions such as sharing, exports and deletions (with a keyed hash of the IP address, not the address), AI usage and cost per workspace (for limits and billing), and content-free routing records described below.

On your device. Essential cookies (see Cookies) and your display preferences in your browser's local storage.

• To run the service: read your papers, extract and check references, answer your questions and run the research you start.
• To keep it safe and working: sign-in, rate limits, abuse prevention, debugging and security investigations.
• To bill you and to answer your support requests.
• To email you about your account and your papers (sign-in links, invitations, “your analysis is ready”). You can turn paper notifications off in Settings.
• To meet legal obligations.

We do not sell your data, show you ads, or build profiles of you for anyone else.

Your uploads, extracted text, highlights, notes, chats, questions, research runs, knowledge bases and usage are never used to train or improve AI models. Not by us and not by the AI providers we use. We do not sell or license your data for training either.

To do the work you ask for, we send the parts of a paper a task needs and your question or instruction to an AI provider (Anthropic, OpenAI) or to our OCR provider (Mistral). They process the request to return an answer, under business terms that forbid training on our customers' data. We turn off response storage where a provider offers it (OpenAI). Providers may keep a request for up to 30 days only to detect abuse, then delete it. The full list is on the subprocessors page.

To make the Auto model choice cheaper and more accurate we record how each request was routed: the kind of task, the model, which checks passed, cost, time and outcome. Never your text, titles, questions, name or email. We read these records to tune our routing rules; they are not used to train any model. A workspace owner can turn them off in Settings, and they are deleted after 180 days.

When an answer fails one of our quality checks, or you report it, we keep a short excerpt: a few sentences around the problem. It is encrypted, readable only by our staff to find and fix the cause, deleted after 180 days (or 30 days after we close the issue, if sooner) and never used to train models. Workspaces that turn off routing records get no automatic incidents.

If we ever want to use customer data to improve a model, we will ask you first and only use the data of people who say yes.

• Members of your workspace see the workspace's papers and work, as their role allows.
• Our subprocessors (hosting, database, AI, OCR, email, payments) handle data only to provide the service, under contract. See the subprocessors page for who they are, what they do and where.
• Services you connect. If you add a source (a connector) or give an outside tool access to a knowledge base, the data you choose to share goes to that service under its own terms.
• Public scholarly databases receive search terms and identifiers (such as DOIs and titles) when we look works up, never your files, notes, chats or identity.
• Authorities, when the law requires it. We push back on requests that are not legally valid.
• A buyer, if the business is sold, bound by this policy.

• Encryption in transit (HTTPS) and at rest.
• Workspace isolation in the database itself: row-level security is enforced on every table that holds private data, so one workspace's data can't be read from another even if the application has a bug. The app connects with a database role that is subject to these rules.
• Files are kept in a private store with a separate area per workspace, and are only reachable through short-lived links issued after an access check.
• Passwords and keys for sources (connectors) are encrypted with a key per workspace.
• Passwords are hashed with Argon2; you can end every session at once from Settings.
• Logs are scrubbed of paper text, prompts, answers, emails and keys; sensitive actions are recorded in an audit log.

No system is perfectly secure. If a breach affects your personal data we will tell you and the relevant authorities as the law requires.

• Your account and workspace data: while your account is open.
• A paper you delete goes to Recently deleted for 7 days, then is removed for good.
• Deleting a workspace or your account hides it at once and removes it for good after 30 days, files included. Until then you can undo it. Papers you added to a workspace other people still use stay in that workspace, without your name on them.
• Exports are kept for 7 days, then deleted. Routing records are deleted after 180 days. Answer-quality excerpts are deleted after 180 days, or 30 days after we close the issue if sooner.
• Database backups keep deleted data for up to 60 days before it is gone everywhere.
• Security audit entries (what action was taken, by which account, never content) are kept after deletion, to protect the service.
• Public scholarly material is kept, because it is not personal to you.

You can, at any time:

• See and export your data: Settings → Workspace → Export gives workspace owners and admins a zip of the workspace's papers, notes, chats, highlights, knowledge bases and more. Anyone else can ask us for a copy of their data.
• Correct it: edit your profile and your work, or ask us.
• Delete it: delete papers, a workspace or your whole account in Settings.
• Object or ask us to limit how we use your data, and opt out of routing records.

For anything you can't do in the app, email support@citationmonster.com. We answer within 30 days. Depending on where you live you may also complain to your data protection authority.

We use only essential cookies: one that keeps you signed in, the ones the sign-in process needs (security check and where to return after signing in), and one that remembers which workspace you have open. We use no analytics, advertising or tracking cookies, and load no third-party trackers. Because these cookies are essential, we don't ask for consent. When you pay, Stripe's checkout page (on Stripe's site) sets its own cookies for fraud prevention.

Our servers and most of our subprocessors are in the United States; our OCR provider is in the European Union. When we transfer personal data out of the EU, UK or Switzerland we rely on the safeguards the law provides, such as the European Commission's standard contractual clauses.

Citation Monster is not meant for anyone under 16. We don't knowingly collect their data; if you think a child has given us data, email support@citationmonster.com and we will delete it.

Email support@citationmonster.com for privacy questions and requests. Organisations that need a data processing agreement can read our DPA summary.

When we change this policy we update the date at the top. For changes that matter, such as a new use of your data, we tell you by email or in the app before they take effect. We will never weaken the promise not to train on your data without your opt-in.