Privacy Policy
Version 2026-10-05. PaperLM ("we") is an invite-only research tool for reading and reviewing machine-learning papers. This policy says what we collect, why, where it goes, how long we keep it and how you control it.
PaperLM is operated by Cheshire Cat Ltd, a company registered in the United Kingdom, which is the controller of your personal data. Questions and requests: [email protected].
What we collect
Your account. Your email address and sign-in method (Google or email and password, through Firebase Authentication), your plan, the invite code you signed up with, the versions of these documents you accepted and when, and when you were last active.
Your Gemini API key. Encrypted with Google Cloud KMS and stored apart from everything else; we keep its last four characters to show you which key is saved. Exactly how it is stored, used and deleted is on Your Gemini API key.
What you put in and what we generate for you. The paper links and PDFs you submit; the reviews, critiques, roasts, images, comics, social posts and other results we generate; your chats, notes, edits and workspaces.
How you use the product. Usage events — for example that you started an analysis, that it finished or failed and at which stage, that you viewed a review, or that a limit stopped you — each with your account id, plan and whether your own key was used. A usage ledger records each model call made for you: which model, how many tokens and the estimated cost. Neither contains the text of your papers or reviews.
Feedback you send. The message, its category, your email address (so we can reply), the page you sent it from, your browser details (user agent, language, time zone, window size), your plan and whether you use your own key. When you send it from a paper or a workspace, also that paper's or workspace's state — its status, current stage, error message and run id — unless you untick "Include the state of this paper" (or workspace). Never the paper's text, its review or your chats.
Server logs. Our servers log requests, including your IP address. Google Cloud keeps these logs for 30 days.
The waitlist. If you joined the waitlist: your email, what you told us, referral details and a salted hash of your IP address — never the address itself.
We use no advertising or analytics cookies and no third-party trackers. Your browser keeps some things locally, in its own storage: your sign-in session, and your display preferences and drafts in local storage — the colour theme, how the library and a paper are shown, chat options, which panels are open, and the settings of a new analysis you have not submitted yet. They stay in your browser; clearing its site data removes them.
Who processes it
- Google Cloud and Firebase host PaperLM: sign-in, the database, file storage, servers, task queues, key encryption (Cloud KMS), logs and our product-metrics warehouse (BigQuery). The database, servers, task queues and key encryption run in Google Cloud's europe-west1 region (Belgium), and the metrics warehouse is in the EU multi-region. File storage — the PDFs you upload and the images and comics we generate — is in Google Cloud's us-central1 region (Iowa, USA). The website's static files are served from Google's global content network.
- Google Gemini API. To analyse a paper, chat or build a workspace result, we send the paper and your instructions to Google's Gemini API with your API key — or, on a plan that includes credits and when you have not saved a key, with ours. How Google handles that content is governed by the Gemini API terms for the key's tier, and the tier matters: on Google's paid tier Google does not use your content to improve its products; on its free tier it may, and human reviewers may read it. Use a key from a Google Cloud project with billing enabled. Google may process this content in any country where it has facilities.
- Google Search. In a paper analysis, the review, critique and roast stages let Gemini search the web with Google Search. Chat about a paper and workspace chat let Gemini search the web and read web pages linked in your messages; both are on by default and can be switched off in the chat. Deep Research in workspaces searches the web by design.
We do not sell your data and do not share it with anyone else.
Transfers outside the EEA and the UK
Google LLC is based in the United States. Your uploaded files and generated images are stored there, and Google may process the other data described here, and the content sent to the Gemini API, outside the European Economic Area and the United Kingdom. Google handles the data we store with it under its Cloud Data Processing Addendum, which includes standard contractual clauses for such transfers. Content sent to the Gemini API with your own key is covered by the Gemini API terms you accepted with Google.
Why (legal bases)
- To provide the service you signed up for (contract): your account, your papers and everything generated for them, your key.
- Our legitimate interests: keeping the service secure and working (logs, limits, abuse prevention) and understanding how it is used so we can improve it (usage events and the usage ledger).
You can object to the processing we base on legitimate interests; write to the address above. We do not make decisions about you by automated means that have legal or similar effects.
How long we keep it
| Data | Kept for |
|---|---|
| Usage events | 90 days |
| Usage ledger (per-call tokens and estimated cost) | 400 days |
| Papers in Trash | 30 days, then deleted automatically |
| Feedback you send | 400 days |
| Server logs | 30 days |
| Waitlist entries | until you ask us to remove yours |
| Your API key | until you delete it, or your account |
| Everything else in your account | until you delete it, or your account |
When you delete your account, your API key is deleted at once and everything else is deleted after 14 days, when the cancellation window ends: papers, results, chats, notes, workspaces, feedback, usage events and the usage ledger, including their copies in our metrics warehouse, and finally your sign-in record. Daily totals that carry no account id are kept. Our database backups may still contain deleted data for up to 98 days, until they expire.
Your rights
You can see, correct, download and delete your data, object to or restrict our use of it, and complain to the data protection authority where you live or work. We answer requests within one month.
- Download your data: Settings → Your data → Download. You get one JSON file with your account details, papers, results, chats, notes, workspaces and feedback. It holds text: images and comics are not in it — save them from the paper or workspace page.
- Delete your account: Settings → Your data → Delete account. You have 14 days to change your mind; cancelling restores your content but not your API key.
- Delete your API key: Settings → Gemini API key → Remove.
- Anything else — correction, objection, a question: write to [email protected].
Children
PaperLM is for adults: you must be 18 or older. (Google requires this of everyone who uses the Gemini API.)
Changes
When this policy changes we update its version above.