Privacy policy

This policy explains how ResearchRadar handles information when you use the ResearchRadar website, web app, or browser extension. It covers your account, saved paper metadata, followed drafts, documents you upload, snapshots that you explicitly ask us to capture, AI analysis of those documents, and subscriptions.

ResearchRadar is designed for researchers and R&D teams. The extension does not provide medical, financial, credit, or advertising services.

Account information. When you sign in or connect the extension, we process the information needed to run your account, such as your email address, account identifier, display name and organisation, and your ORCID iD if you sign in with ORCID.

Waitlist sign-ups. If you join the waitlist, we process the email address and role you give and, if you add them, a note and a price band, together with the IP address, browser user agent, referring page and campaign tags of the sign-up request. We use your email address to confirm your sign-up, to tell you it is confirmed, and to send one early-access announcement, which carries an unsubscribe link.

Paper and document metadata. When you save or follow a document, we process information such as its title, source URL, authors, DOI, publication date, editor identity, and selected project or tab identity.

User-requested source content. When you choose Capture snapshot, we process the saved Overleaf source archive, a paper PDF, or the saved plain text of the selected Google Docs tab. Source captures are limited to 2 MB and paper PDFs to 25 MB. Unsaved editor buffers are not captured. For a captured paper PDF we also record the licence the publisher declares for it, when one can be found.

Documents you upload. In the web app you can upload a PDF or Markdown document of up to 25 MB.

Analysis results. When AI analysis runs, we store the text extracted from the document and the claims, evidence, and topics derived from it, linked to your account.

Billing information. If you subscribe, Stripe collects your payment details; we do not receive your full card number. We store the Stripe customer and subscription identifiers, your plan and its status, and when you completed checkout.

Extension settings. The extension stores connection preferences, connection status, and a stable browser-installation identifier in browser storage. It does not store passwords, cookie values, access tokens, or snapshot bytes in browser storage.

Connected browsers. While the extension is connected, it sends that random installation identifier, the browser name, and the extension version to ResearchRadar about every five minutes, so the web app can show which browsers are connected to your account. An entry expires after seven days without a check-in, and disconnecting removes it.

We use this information to run your account, connect the extension to it, save and display paper and document metadata, store and retrieve the documents and snapshots you added, prevent duplicate snapshots, provide downloads, analyse documents when you have allowed AI analysis, manage your subscription, and let you remove saved documents and their snapshots.

Weekly digest. If monitoring is on for at least one of your radars, we email you once a week, on the day and at the time you choose, a list of the new papers your radars found; no email is sent in a week without new papers. We keep a record of each digest and the papers it listed so that a paper is not sent to you twice; it is deleted with your account. You can turn the digest off in your profile or with the unsubscribe link in every digest, which does not affect sign-in or billing emails.

Detection on a supported page does not by itself upload document contents. Source content is sent only after you choose an action such as Capture snapshot or Add to graph.

The extension does not use this information for advertising, credit decisions, lending decisions, or automated background document uploads.

ResearchRadar can analyse the documents you upload or capture to extract their claims, evidence, and topics. Analysis runs only after you allow it: you are asked before your first document is analysed, and you can turn it off at any time in your profile. Turning it off stops analyses that have not finished; text that was already sent to the model provider cannot be recalled.

To analyse a document, we send its extracted text to an AI model provider: OpenAI or Google (Gemini API), depending on the model in use. The provider processes the text and returns the analysis to us. For OpenAI requests we ask OpenAI not to store the response for its own products; any retention for abuse monitoring is governed by the provider's terms.

When you save a published paper that has a DOI, we look up its public record in OpenAlex and Crossref and may assess its technology readiness. That assessment sends the paper's public metadata, not your documents or account details, to an AI model provider (Anthropic, OpenAI, or Google, depending on how the service is configured), and may run web searches derived from that metadata through Exa.

If you turn on monitoring for a radar, the interest keywords you enter for it are sent to OpenAlex as a search query to find new papers that match them.

activeTab and scripting are used to inspect the page you choose and to request a bounded export from supported paper, Overleaf, or Google Docs pages. storage is used for local connection state. alarms are used for connection and sign-in status checks and browser check-ins. Host access is limited to the ResearchRadar service, supported research and editor sites, and the OpenAlex API.

The extension includes its executable code in the package. It does not load or execute remotely hosted JavaScript, WebAssembly, or other remote executable code.

We do not sell user data and do not use it for advertising. Saved metadata, documents, and snapshots are sent to ResearchRadar's API and private storage so that they can be associated with your account and made available to you.

These service providers process data on our behalf, only to provide the service described next to each:

• Railway: hosts the ResearchRadar API, database, and private file storage.
• Vercel: hosts the website and the web app.
• OpenAI or Google (Gemini API): AI analysis of documents you have allowed us to analyse; which one is used depends on the model in use.
• Anthropic, OpenAI, or Google: technology-readiness assessment of published papers' public metadata; which one is used depends on the service configuration.
• ORCID: sign-in with your ORCID iD, if you choose to sign in that way.
• Exa: web searches derived from the public metadata of published papers.
• Stripe: payments and subscription management.
• Resend or Twilio SendGrid: delivery of sign-in codes, account emails and waitlist emails.
• Sentry: error monitoring, configured not to collect default personal data.

We also query the public scholarly databases OpenAlex and Crossref with a paper's DOI, and OpenAlex with the interest keywords of radars you monitor. To show a paper's details, the extension may look up its DOI with OpenAlex directly from your browser.

When you capture an Overleaf or Google Docs source, the export request uses your browser's existing session with that editor. The extension does not send your editor cookies or credentials to ResearchRadar.

The use of information received from the extension will adhere to the Chrome Web Store User Data Policy, including the Limited Use requirements.

We retain account-associated metadata, documents, snapshots, and analysis results while they are needed to provide the service or until you remove them.

Choosing Remove deletes the document's stored files and stops any analysis that has not finished. The analysis records derived from it, such as extracted text, claims, and evidence, are hidden at once and permanently deleted after a 7-day grace period.

You can delete your whole account yourself under Your data in your profile in the web app. Deleting the account permanently removes your documents, snapshots, uploads, radars, analysis results, and subscription, and signs you out on every device.

Some records are not removed by account deletion: billing records such as invoices and payments that Stripe keeps on its side, a log entry recording that a deletion took place (without your name, email address, or ORCID iD), your bare account identifier in the audit trail of any complimentary access you were given, and delivery logs that our email provider may keep for emails already sent.

Disconnecting the extension stops extension session use but does not delete website records. To ask about data retained in operational backups, contact us at hello@researchradar.dev. Backup copies may remain temporarily until the applicable backup cycle expires.

You can choose whether to save a paper, follow a draft, upload a document, or capture source content, and whether ResearchRadar may analyse your documents with AI. Uploading a document in the web app needs that permission, because every uploaded document is analysed. Saving papers and capturing snapshots with the extension does not; those snapshots are analysed only after you allow it. You can withdraw the permission in your profile, disconnect the extension, and remove saved documents and snapshots.

Under Your data in your profile you can download a copy of the data ResearchRadar stores about your account and delete your account. You can also contact us to request access, correction, or deletion of personal information, subject to applicable law.

We use only the cookies the service needs to work: a session cookie that keeps you signed in, a short-lived oauth_state cookie set by the ResearchRadar API while you sign in with ORCID or connect an AI client, which protects that sign-in from forgery, and in the web app a cookie that remembers whether the sidebar is open. The website and web app also keep your light or dark theme choice and unfinished forms in your browser's storage.

We do not use analytics or advertising cookies.

We may update this policy when the service or the extension changes. The effective date below will be updated when a material change is published.