Privacy Policy
Effective 1 August 2026
Graspr reads web pages. That deserves a straight answer rather than a page of hedging, so here is exactly what happens.
What Graspr reads, and when
Graspr reads the page you are actively viewing, and only when you open it on that page — by clicking the Graspr icon, using the keyboard shortcut, or setting a per-site rule yourself.
A lightweight script is present on pages so the sidebar can open instantly. It does not read page content or make any network request until you open Graspr for that page.
When you do open it, we send a topic and URL derived from the page to the services listed below, to find discussion about it. In reader mode we also process the article's text so it can be displayed cleanly.
What we never do
- We never sell or rent your data. To anyone, for any price.
- We never use your data for advertising, or to build a profile of you.
- We never send your browsing history. Only the page you actively open Graspr on.
- We never send page content, URLs, titles, highlights or anything you type to our analytics.
Who receives what
| Service | What it receives | Why |
|---|---|---|
| Serper | Search queries derived from the page you're viewing | Finds articles and videos about the page |
| Google (Gemini embeddings) | The page title and candidate discussion titles | Ranks results by relevance to the page |
| Hacker News, Reddit, GitHub, Bluesky, X, Dev.to | A topic or URL derived from the page | Finds existing discussion about it |
| Clerk | Your email address and login credentials | Accounts and sign-in |
| Supabase | Your saved articles, highlights and settings | Storage, if you're signed in |
| Dodo Payments | Billing details you enter at checkout | Subscription payments (we never see your card) |
| Notion / Readwise | Articles and highlights you save | Only if you connect them yourself |
| PostHog | Anonymous usage events — never page content | Product analytics; can be turned off |
| Sentry | Error reports, scrubbed of URLs and personal data | Diagnosing crashes |
Ranking, and the on-device option
By default, ranking results by relevance runs on our server: the page title and the candidate discussion titles are sent to Google's embedding API. No article body and no personal data is included.
If you would rather nothing about the page leave your browser, turn on Rank on this device in Settings → AI. Graspr then downloads a small model (~25MB, once) and does the ranking locally. It is slower and somewhat less accurate — that is the trade.
Analytics
If enabled, we record anonymous usage events: which source tabs you open, how many results came back, how fast, and whether a page failed to load in the reader. Each install gets a random identifier that is not linked to you unless you sign in.
We do not record page URLs, titles, content, your highlights, your notes, or your search queries. Turn it off any time in Settings → Privacy; it takes effect immediately and discards anything not yet sent.
What is stored, and where
On your device: preferences, per-site rules, reading positions and cached results.
On our servers, only if you are signed in: articles you save, your highlights and notes, your account email, and your subscription status. Connected Notion and Readwise tokens are encrypted at rest.
Signed out, Graspr stores nothing about you on our servers.
Your data, your call
- Export everything — one button on your dashboard.
- Delete everything — one button on your dashboard. It is immediate and permanent.
- Disconnect Notion or Readwise — the stored token is deleted.
- Uninstall — removes everything stored locally.
Children
Graspr is not directed at children under 13, and we do not knowingly collect their data.
Changes
If this policy changes materially, we will update the effective date above and tell signed-in users by email. We will not quietly broaden what we collect.
Contact
Questions, or a privacy request: get in touch. Privacy requests are answered within 30 days.