Papero.ai Privacy Policy

Types of information We collect, Why and How

We will only collect personal data from you that we consider to be necessary in the context and purpose in which it's given.

Through your use of our services we may also collect personal data from you about someone else. If you provide us with personal data about someone else, you must ensure that you are authorised to disclose that data to us and that, without us taking any further steps required by applicable data protection or privacy laws, we may collect, use and disclose such personal data for the purposes described in this Policy.

You must, therefore, take reasonable steps to ensure that the third party concerned is aware of and consents to the various matters detailed in this Policy, including: the fact that their personal data is being collected; the purposes for which that data is being collected; the intended recipients of that data; and the third party's right to obtain access to the data (including details of how to request access). Where requested to do so, you must assist us with any requests by the third party to access or update the personal data you have collected from them and provided in connection with our Services.

How to access and control your data

You can review, edit or delete your personal data by contacting us by email to contact@papero.ai. We will respond to any request to access or delete your personal data as soon as possible, but certainly within 30 days.

User Information

When you register for an Account (as defined in the Terms of Service), we collect your e-mail address. Every Account, on the Free plan too, has a customer record at Stripe, our payment provider, with its e-mail address and user name (and the referral code, if you signed up through one). The Free plan needs no billing information. The Plus plan needs billing information: Stripe processes the card, and we keep a VAT record of the purchase with your billing name and address, the company name and VAT number if you give them, the brand, country and last 4 digits of the card, and the IP address you bought from.

When you register for an Account, we collect the e-mail address and any user profile fields which are filled out, such as the user's name. This information is used to identify you in connection with the Service, for billing as described above, and for email communication between you and the support staff of the site you are using.

Before Papero.ai opened to everyone, people could join a waitlist without creating an Account. The waitlist form collected only the email address and the language of the page used to sign up, and we emailed a link to confirm the address. An unconfirmed address was used for nothing else; a confirmed address was used only to email its owner when Papero opened. To have your address removed from the waitlist, write to contact@papero.ai.

The newsletter forms at the bottom of papero.ai's pages and on the Papero blog (papero.ai/blog) collect only your email address, the language of the page and that you agreed to receive our emails. We first email you a link to confirm it, and until you confirm we send you nothing else. Once you confirm, we email you new blog articles and news about Papero. The list is kept by MailerLite, which sends these emails for us. Every email has an unsubscribe link that stops them, or write to contact@papero.ai.

When you confirm your email address, or sign up with Google or GitHub, your Account's email address joins our list for emails about getting started with Papero, with whether you are on Free or Plus and, when you confirmed by email, the language of the page. MailerLite keeps the list and sends these emails for us. Every email has an unsubscribe link, and deleting your Account takes you off the list.

If you ask us for help, a member of the Papero team may view your account as you see it, read-only and for up to 15 minutes at a time, only to answer your request. Every such view is logged.

Cookies

At Papero.ai, we are committed to protecting the privacy and security of our customers and website visitors. The Papero.ai team members are customers themselves, both of Papero.ai and other Internet sites. We therefore fully appreciate and respect the importance of data privacy and security on the Internet.

We believe you should feel both informed and empowered when it comes to our handling and usage of your information.

A cookie is a small amount of data which often includes an anonymous unique identifier that is sent to your browser from a web site's computers and is stored on your computer's hard drive.

Cookies are required to use our site in order to uniquely identify your browser and user preferences while logged in. Our session cookie lasts up to 7 days, after which you will be required to log in again.

We use the following types of cookies on our website:

• Strictly necessary cookies. These cookies are essential for you to browse our websites and use the features.
• Performance cookies. These cookies collect information about how you use our websites. This data may be used to help optimise our website and make it easier for you to navigate.
• Functional cookies. These cookies, and similar data stored in your browser, allow our websites to remember choices you make, such as your answer to our cookie banner.
• Third Party cookies. Third party cookies are those placed by parties other than Papero.ai. On our website these are Google Analytics and Hotjar, and they are set only after you accept them. They are subject to the privacy policies of Google and Hotjar. The support chat by Crisp sets its cookie only when you open the chat (see Support chat below). The sign-up form uses Google reCAPTCHA to stop automated sign-ups. It loads with the sign-up page, because the form needs it, and Google may set its own cookies and receive information about your browser and how you use the page, under Google's privacy policy.
• Analytics cookies. We use Google Analytics and Hotjar to understand things like how long a visitor stays on our website, what pages they find most useful, and how they arrived. Neither sets a cookie until you click Accept in our cookie banner: until then Hotjar does not load, and Google Analytics, in Google's consent mode, sends only cookieless measurement signals. You can change your answer at any time with the Cookie settings link at the bottom of every page. We also use Google Tag Manager, a Google service that loads the measurement tags we set up in it; we use it only for the same kind of analytics, never for advertising. It also stays off until you click Accept, and it sets no cookies of its own. Google explains how it uses this data at policies.google.com/technologies/partner-sites.

Advertising

Papero.ai does not use advertising cookies, tracking pixels or interest-based advertising, and we do not sell your personal data or share it with advertisers.

When you accept cookies, Google Analytics runs for measurement only: we never give it consent to use your visit for advertising.

Page speed measurement

Every page of papero.ai and of the Papero app loads a small script from speed.loadfocus.com, LoadFocus Page Speed, which LoadFocus LTD (the company behind Papero) runs. When you leave a page, the script sends once the page's path (without its query string; for a shared document link, a placeholder instead of the link), how fast the page loaded and responded, and the type of device and network connection. It sets no cookie and stores nothing in your browser. Your IP address reaches LoadFocus with that request: it is used only to find your country and to limit abuse, and the address itself is never stored; for those limits LoadFocus keeps a scrambled code of the network it belongs to for up to 40 days. The country and the type of browser and device are counted in totals with the timings, never per visitor. We do this on the basis of our legitimate interest in keeping the site and the app fast.

Data Storage

Papero.ai uses third party vendors and hosting partners to provide the necessary hardware, software, networking, storage, and related technology required to run Papero.ai and its websites and services. Although Papero.ai owns the code, databases, and all rights to the Papero.ai application, you retain all rights to your data.

Security

Papero.ai is hosted at Amazon Web Services (https://aws.amazon.com). For more information on their servers and security, please see Amazon Web Services's security white paper.

Any personal data transferred to us over the internet is protected with Secure Sockets Layer / Transport Layer Security (SSL/TLS) Encryption Certificate technology. This ensures that all personal data transferred is encrypted to prevent eavesdropping and tampering.

Our services can be accessed through the use of an individual user login and password. To protect the confidentiality of data, you must keep your password confidential and not disclose it to any other person. Please alert us immediately if you believe your password has been misused. Additionally, always logout and close your browser when you finish your session, especially if you're on a public computer. You are ultimately responsible for administering and safeguarding any passwords or memorable words created to control access. Please note, we will never ask you to disclose your password in an unsolicited phone call or email.

If you have any questions about the security of your data, you can contact us at contact@papero.ai.

Your documents and the AI features

Your documents are private to your Account, unless you share a link to one (see Sharing a document for comments below). When you use an AI feature (autocomplete, rewrite, outlines, section drafts, chat, Proofread, Claims, Peer review, Tone of voice, the Citation report, chart generation and equation generation), the text it needs is sent to OpenAI, which generates the answer in the United States. OpenAI does not use that text to train its models and keeps it for up to 30 days to monitor abuse. We do not use your documents to train AI models. When you use Describe in the chart panel, Papero sends OpenAI your description of the chart and, when you are editing an existing chart, the chart's current source (Plotly JSON or Mermaid text, up to 8,000 characters), and nothing else from the document. The chart's source and its picture are stored in your document like the rest of its content. When you use Describe in the equation editor, Papero sends OpenAI your description of the equation and, when you are editing an existing equation, its current TeX (up to 4,000 characters), and nothing else from the document. The equation's TeX is stored in your document like the rest of its content. To find papers, Papero sends your search words, or the sentence you ask it to find a citation for, to OpenAlex; to look up a paper by its DOI, PMID, arXiv ID or ISBN, it sends that identifier to OpenAlex, doi.org and Crossref (or DataCite, for DOIs registered there). On Plus, when a suggestion or a section draft is set to cite papers, Papero also sends OpenAlex the section headings, the section prompts and key terms from the document prompt, to find papers for that section, and the request to OpenAI includes the titles of your own sources so that the answer can cite them. Papero checks every citation on its own server before it is shown. When you run Claims, Papero sends OpenAlex short search queries built from the document's claims. When you run the Citation report (Plus), the request to OpenAI includes each cited sentence (up to 40), the titles of the sources it cites and those sources' abstracts from OpenAlex. The paper-lookup services (OpenAlex, doi.org, Crossref and DataCite) receive only the search words, the Claims search queries, the section headings, prompts and terms described above, or the identifier, and Papero's contact email address (in the request, so the services can reach Papero), never your name, your email address or anything else about your Account. To check the sources a document cites, Papero sends OpenAlex only those works' DOIs or OpenAlex ids and the same contact email address. When a suggestion, a section draft, the Citation report or a chat answer uses a paper whose PDF you uploaded, the request to OpenAI also includes passages of that PDF's text. When you ask a question in chat, Papero sends OpenAI your question, the earlier messages of that conversation, passages of your uploaded PDFs that match it, the authors, years and (on Plus) titles of the papers they come from and, when the conversation includes the current document, up to about 12,000 characters of that document's text. On Plus, chat can also use the titles and abstracts of papers in your library that have no PDF. Papero fetches those abstracts from OpenAlex (sending only each work's DOI or OpenAlex id) and sends the titles and abstracts of up to 12 of your papers to OpenAI to find the ones that match your question and to answer from them.

When you sign up, the app asks for your role, your field of study, what you are writing first and where you heard about Papero; Papero sends your role and field to OpenAI along with your text so that the writing is pitched to you, uses the source only to understand how people find Papero, lets you change your role and field in Settings, and deletes all four answers with your Account.

When you import your references from Zotero, Papero sends the API key you enter to api.zotero.org to read your library. If you tick "Remember my key", Papero keeps the key encrypted with AWS Key Management Service and uses it only when you import from Zotero again; it deletes the key when you disconnect Zotero, when Zotero refuses it the next time you import, when you have not used it for 12 months, or when you delete your account. Without the tick, Papero sends the key only while that import runs and does not keep it. For each reference you import this way, Papero also keeps your Zotero user ID and that reference's Zotero item key with the source, so that a Word export on Plus can link each citation to the same item in your own Zotero library. They stay with the source after you disconnect Zotero, and are deleted when you delete the source or your account.

When you rate an AI suggestion or a Chat answer with a thumbs up or down, Papero records the rating, any reasons you tick and, if you write one, your comment (up to 500 characters), together with your Account and the AI model that wrote the suggestion or answer. We use this only to improve Papero's suggestions and answers. Your comment is never sent to OpenAI or anyone else. The suggested text is not stored with your feedback, and a rating of a Chat answer stores only which answer it was: the answer stays in its conversation (see Chat with your library), and deleting the conversation does not delete the rating. We keep feedback for 12 months, and delete it sooner if you delete your Account.

When you import from Mendeley, you sign in at Mendeley, so Papero never sees your Mendeley password, and it discards the access Mendeley gives it as soon as the import is done.

PDF files you upload

In the Papero app you can upload PDF files to your library. Papero stores each file privately with your Account, at Amazon Web Services. Your browser extracts the text of each file and sends it to Papero, which stores it with your Account as passages. Your files and their text are never shared with other users and never used to train AI models. When you use an AI feature with a paper whose PDF you uploaded (a suggestion, a section draft, the Citation report or a chat answer), passages of its text are sent to OpenAI for that request. To find the passages that match what you are writing or asking, Papero also sends passages to OpenAI to compute embeddings (lists of numbers that describe what a passage is about), which it stores with the passages. OpenAI does not use this text to train its models and keeps it for up to 30 days to monitor abuse. Papero deletes a PDF file and its text when you remove the PDF or its paper, when you delete your Account, or 12 months after your last sign-in (in that case we email you 30 days before). Your library entries and the citations in your documents stay. Backup copies are deleted within 30 days.

Chat with your library

In the Papero app you can ask questions about the PDFs in your library and the document you are writing. Papero stores each conversation (your questions, the answers and which of your sources each answer cites) privately with your Account so that you can return to it. Answers can quote passages of your PDFs; a quote stays in its conversation after you remove the PDF, until the conversation is deleted. On Plus, chat can also use the titles and abstracts of papers in your library that have no PDF. Papero fetches those abstracts from OpenAlex (sending only each work's DOI or OpenAlex id) and sends the titles and abstracts of up to 12 of your papers to OpenAI to find the ones that match your question and to answer from them. Conversations are never shared with other users and never used to train AI models. If you ask us for help and a member of the Papero team views your Account read-only, they can also read your conversations, but they cannot change or delete them. Papero deletes a conversation when you delete it, when you delete the document it belongs to, when you delete your Account, or 12 months after its last message. Papero keeps your 100 most recent conversations and deletes older ones. Backup copies are deleted within 30 days. Your data export includes your conversations.

Workflows

When you run a Workflow, your topic, the titles and abstracts of the papers it finds, and passages of your own PDFs it uses are sent to OpenAI to plan and write the document, as for chat. Papero also sends OpenAlex short search queries that the AI model writes from your topic, to find papers. Like other paper searches, these go to OpenAlex with only Papero's contact email address, never your name, your email address or anything else about your Account. Papero stores each run (your topic, its settings, the sections it wrote and the papers it cited) privately with your Account so that you can return to it. Runs are never shared with other users and never used to train AI models. Papero deletes your runs when you delete your Account, and backup copies are deleted within 30 days. Your data export includes your runs.

Images you add to documents

In the Papero app you can add images to your documents. Papero stores each image privately with your Account, at Amazon Web Services. Your browser removes the location and camera data (EXIF and GPS) from an image before it is uploaded. An image is shown only to its owner and never to other users, except in a document its owner shares by link, where readers of that link can see its images (see Sharing a document for comments). Images are never sent to AI models. Papero deletes your images when you delete your Account, and also 30 days after no document and no saved version of your documents uses an image any more.

Sharing a document for comments

In the Papero app the writer of a document can create a link to that one document. Before creating a link, the writer sets a display name, and everyone who holds the link sees it. Anyone who is signed in to Papero and has the link can read the document (its title, text, figures and formatted reference list; tables, equations and charts are shared as part of the text) and comment on passages of it. They cannot edit it. They see the document as it is when they open it, including changes the writer makes after sharing. They do not see the writer's library, other documents, the document's prompt, outline or saved versions, or the writer's chat conversations. The writer can stop the link, or reset it to get a new link, and the old link then stops working at once. Archiving a shared document pauses its link, and restoring the document makes the same link work again; stop or reset the link to end it. A figure a reader has already loaded may stay viewable in their browser for up to about two hours after the link is stopped or reset. While a link works, everyone who has it can read the document, so share a link only with people you want to read it.

A comment and the display name of whoever wrote it are visible to the writer and to everyone who holds the same link. The writer also sees the email address of each commenter. A link that has been reset starts with no comments visible to the people who hold it (the writer still sees all of them). To comment, you need a confirmed email address and a display name, which Papero stores in your Account's profile. The writer can get an email about new comments, at most one email an hour for each link, and only if the writer's email address is confirmed. It names the document and how many comments are new, never their text. It is sent through Amazon SES (already listed as a sub-processor) and can be turned off in Settings. Sharing adds no new processor, cookie or tracker.

You can delete your own comments, and the writer can delete any comment on their document. The text of a deleted comment is erased at once. Other comments are kept for as long as the document exists. Deleting a document deletes its link and its comments. Deleting a writer's Account deletes the links and comments on their documents, and deleting a commenter's Account deletes the comments that person wrote. A comment that others have replied to stays, shown as “Comment deleted” with no name. The data export includes your links (without the secret part of the link) and the comments you wrote or that were left on your documents. Backup copies are deleted within 30 days.

Papero Web Importer for Chrome

Papero Web Importer is an optional Chrome extension that saves papers and web pages to your Papero library. It works only with a Papero Account and your existing papero.ai session in the browser, and it contacts only app.papero.ai. Its "Add to Papero" buttons are on by default. While they are on, on each page you open (except Papero's own sites) it reads, in your browser, the page's citation and Open Graph meta tags, the page title and the URL, and the addresses and visible text of the page's links, to find papers (the page's own paper, and links to papers by DOI, arXiv or PubMed) and place a button beside each one, or, for the page's own paper, in the bottom-right corner if it cannot find the page's title. When you open its side panel and these give no DOI for the page, it also looks for a DOI in the page text; that text is never stored or sent. You can switch the buttons off for a site or everywhere; the extension then reads those pages only when you open its side panel or use its right-click menu, where "Add this page to Papero" opens the side panel and "Add this link to Papero" saves the paper a link points to. Nothing leaves your browser until you click a button, use the right-click menu or open the side panel. For the page you are on, the extension sends app.papero.ai only the fields it extracted (an identifier such as a DOI, arXiv ID or PubMed ID, the title, the URL and the site name), to look up the reference, to list or create your collections and to save the reference. For a paper link, the only thing it sends from the page is that paper's identifier (a DOI, arXiv ID or PubMed ID), plus the link's text if the paper's record has no title.

The extension keeps only its settings in your Chrome profile (the theme, the "Add to Papero" switch and your choices for each site, where quick saves go (the last collection you used, or always My Library), and the collection you picked last), where Chrome Sync may sync them. It keeps no password, token or key. It has no analytics, trackers or error-reporting service, it does not read your browsing history or form fields, and nothing it handles is sold or shared. Uninstalling it removes its settings. A reference you save is kept in your library like any other source and stays there until you delete it, even after you uninstall the extension. How we keep and delete what is in your library is set out under Our retention of your information.

Support chat

The support chat on papero.ai is provided by Crisp (Crisp IM SAS, France). Nothing is sent to Crisp until you press the chat button: only then does your browser load the chat from Crisp's servers, which receive your IP address and details of your browser. Crisp then sets a cookie so that your conversation is still there when you come back; it lasts up to 6 months, and while it is there the chat loads on its own so that you see our replies. Crisp stores the messages you write, and an email address if you give one, so that we can reply; we use them only to answer you. To have a conversation deleted, write to contact@papero.ai. Google Analytics records that the chat was opened and that a message was sent, never what it said, and like the rest of Analytics it sets no cookies unless you accept them. Hotjar does not record the chat. In the Papero app, "Chat with us" in the "Help" menu opens the same chat on Crisp's own page, in a new tab; the app itself loads nothing from Crisp.

Connecting Papero to an AI assistant

You can connect Papero to an AI assistant such as Claude or ChatGPT by adding https://app.papero.ai/mcp as a connector there and signing in with your Papero account. You sign in on Papero's own page; the assistant never receives your password or an API key. The assistant can then use a fixed set of read-only tools: search for papers, look up a paper by DOI, PMID, arXiv ID or ISBN, format references in one of nine citation styles, list your documents, read the text and references of a document, and see your plan and allowances. It cannot change or delete anything, and it does not run Papero's AI features. Papero sends nothing to the assistant on its own: the assistant receives the answers to the tools it calls, when you ask it to or when it decides to, and that can include the text and references of a document you ask it to read. A paper search or lookup made this way goes to the same paper services as one made in Papero, as described above, and a search counts against your search allowance. What the assistant's provider (for example Anthropic or OpenAI) does with what it receives is governed by that provider's own terms and privacy policy, not by this Policy. That provider is not a sub-processor of Papero, and Papero is not responsible for how it handles that information. Papero keeps a record of each connection (which app, when it was connected and when it was last used) and the access tokens it issued, and shares the tokens with no one. Papero also records each tool call the assistant makes: which app made it, the tool's name, whether it worked (and, if not, the kind of failure) and when. Papero does not record what the assistant sent to the tool or what the tool answered. If the assistant sends too many requests in a minute, Papero also records, at most once a minute, that it refused them: which app sent them and when. These tool call and refusal records are kept for 400 days, are included in your data export and are deleted with your Account. Access lasts until you disconnect the app under Settings, Connected apps in the Papero app. Access tokens expire after 30 days and are renewed while the app is in use, for up to 90 days; after that you connect the app again. Deleting your Account deletes the connection.

Free tools

The free tools at papero.ai/tools (such as DOI to BibTeX, DOI to citation and PMID to citation) work without an Account. To look up the identifiers you paste (DOIs, PubMed IDs or PMIDs, arXiv IDs and ISBNs), Papero sends them to OpenAlex, doi.org (run by the DOI Foundation), Crossref and DataCite, together with Papero's contact email address. For these lookups the services receive only the identifiers, never your IP address or anything else about you. Some of them are in the United States, and each is an independent recipient, not a processor acting for Papero. Papero does not save the identifiers you paste or link them to you; the records it finds stay in memory for a few minutes, so that an identifier pasted twice is looked up once. When you use the reference checker, Papero's server reads the references you paste, looks for a DOI in each one and compares each reference with the record it finds. For a reference with a DOI, only the DOI is sent on, to the same services. For a reference without a DOI, the reference checker sends part of that reference's text (up to 300 characters, without any DOI or web address) to Crossref's search, and, when it looks like a book, the first author's family name and the main title (or the ISBN) to Open Library, run by the Internet Archive in the United States, to find a matching record. When a record found this way has a DOI, only that DOI is sent on, to the same services. Open Library is also an independent recipient, not a processor acting for Papero. Nothing else about you is sent. The text you paste is not saved or written to any log; as with identifiers, the searches and what they find stay in memory for a few minutes, so that a reference checked twice is searched once. To apply the limits on how many lookups and checks you make in a minute and in a day, Papero turns your IP address into a code with a secret key that changes every time our service restarts, so the code cannot be turned back into your address, and keeps that code for no longer than a day. Your IP address itself is never written down. We do this on the basis of our legitimate interests in providing the tools and preventing abuse. The tools set no cookies of their own.

Disclosure

Papero.ai may disclose personally identifiable information under special circumstances, such as to comply with subpoenas when a users actions may violate the Terms of Service.

Our retention of your information

Papero.ai retains personal data for as long as necessary to provide our services or for other essential purposes such as complying with our legal obligations, resolving disputes and enforcing our agreements. Because these needs can vary for different data types in the context of different services, actual retention periods can vary significantly. When you delete your Account, we delete its documents and their versions, its sources and collections, its chat conversations, its usage records, its sharing links, and the comments left on its documents or written by it. Copies may stay in our backups for up to 30 days before they are deleted. PDF files you uploaded and the text extracted from them are deleted with your Account too, and also when you remove the PDF or its paper, or 12 months after your last sign-in (we email you 30 days before). Your library entries and citations stay, and backup copies of the files and their text are deleted within 30 days.

If Papero.ai needs to alter, restrict processing of your data, or remove your data we will inform you.

Changes to this Privacy Policy

Papero.ai may modify or update this Policy when necessary to reflect feedback and changes in our services. Visiting our website and/or using our services after any modification to this Policy will constitute your acceptance of such modification and updates. When we update this Policy we will revise the "Updated" date at the top of the Policy.

If there are material changes to the Policy or in how Papero.ai uses your personal data, we will notify you either by posting a notice of such changes before they take effect or by directly sending you a notification. We encourage you to regularly review this Policy to learn more about how Papero.ai is using and protecting your information.

Request my data

Contact Papero.ai to request your data and get it sent to your email address. You can do so here.

Contact Us

If you have any questions regarding this Privacy Policy, please contact us through email at contact@papero.ai.

Copyright © 2026 Papero.ai. All rights reserved.