Shakvaro.
ScreenMind

ScreenMind · Extension Privacy

Privacy Policy

The short version: every screenshot, every OCR pass, every search runs locally in your browser. No accounts, no telemetry, no remote servers of ours to send anything to. Last updated June 8, 2026.

1. What ScreenMind is

ScreenMind is a Manifest V3 Chrome extension that captures screenshots, extracts text via local OCR, tags the content, and surfaces context-aware actions. It runs entirely inside your browser. There is no remote ScreenMind server — your captures and the text inside them never leave your device and are never sent to us. We are physically unable to see what you capture.

2. What we don't collect

This is the heart of the policy. ScreenMind does not collect:

  • No account — the extension has no sign-up, login, or user profile.
  • No screenshots, image data, or OCR text sent off your device.
  • No browsing history, URLs you visit, or page contents sent off your device.
  • No personal information (name, email, phone, location, contacts).
  • No advertising identifiers and no third-party analytics or ad SDKs.
  • No selling or sharing of personal data — there is none to sell.

Verified by source inspection: the codebase contains zero fetch or XMLHttpRequest calls to external hosts — only chrome.runtime.getURL for bundled assets. You can audit the source yourself.

3. How it works (the technical truth)

When you capture, the image is held in memory as a blob. A thumbnail and metadata get written to IndexedDB (via Dexie). The blob is then preprocessed — grayscale, contrast, binarize — and tiled vertically for long captures, before being handed to a Tesseract.js workerrunning inside an offscreen document. The offscreen doc exists because Tesseract needs DOM access that the MV3 service worker doesn't have. The OCR output is then run through regex detectors (code / price / contact) to produce a content type and tags. Every step happens on your device.

4. Data stored on your device only

ScreenMind stores the following on your device only. None of it is transmitted to us:

  • IndexedDB (Dexie): screenshots (image blobs + thumbnails), OCR text, tags, collections, leads, prices, tasks.
  • chrome.storage.local: settings, the active mode (Dev / Business / Shopper / Research), the onboarding-complete flag.
  • chrome.storage.session: short-lived hand-off keys such as pendingScreenshotId when opening the side panel after a capture.

Captures are deletable from the side-panel library — individually or as a multi-select batch. Clearing the extension's storage or uninstalling it wipes everything. We hold no copy.

5. Page content we read (and don't read)

ScreenMind reads page content only when you explicitly trigger a capture: the rendered pixels of the visible tab, selected area, full page, or chosen screen/window. It does not read form values, cookies, localStorage, IndexedDB of other sites, or any other page state outside the screenshot itself. It does not log keystrokes. The area-selector overlay is injected only on the active tab when you start an area capture, and is removed when the capture finishes or is cancelled.

6. Third parties

ScreenMind has no backend. Two third parties are inherent to running a Chrome extension:

  • Tesseract.js (bundled).The OCR engine ships inside the extension and runs locally via WebAssembly. No network call is made to load it — it's served via chrome.runtime.getURL.
  • Google Chrome / Web Store. If you install from the Chrome Web Store, Google may collect install and diagnostic metrics under Google's privacy policy. This is standard for any Chrome extension and is outside ScreenMind's control.

7. Permissions we use

  • activeTab, tabs — identify and capture the current tab.
  • scripting — inject the area-selector overlay and the full-page scroll stitcher.
  • offscreen — run the Tesseract OCR worker in an offscreen document (the MV3 service worker has no DOM).
  • desktopCapture— show the native screen/window picker when you choose "Screen / window" capture.
  • storage, unlimitedStorage — store screenshots, thumbnails, OCR text, and settings locally. Screenshots can be large, so the unlimited quota is required.
  • contextMenus — add right-click capture options.
  • sidePanel — power the library UI.
  • host_permissions: <all_urls> — required so capture can work on any site you visit. It does not grant us anything: we never see what you capture.

ScreenMind does not request access to your camera, microphone, geolocation, downloads, identity, history API, cookies, or any other sensitive permission.

8. Children

ScreenMind is suitable for general audiences and does not knowingly collect any personal data from anyone, including children. Because the extension collects no personal data at all, it poses no specific risk to children's privacy.

9. Changes to this policy

If we change how ScreenMind handles data, we'll update this page and revise the date above. Material changes will be reflected in the Chrome Web Store listing and, where relevant, in the extension itself.

10. Open source & license

ScreenMind is open source under the MIT license. Full source is at github.com/shakiltousif/ScreenMind. Anyone is welcome to read it, audit it, fork it, or file an issue.

11. Contact

Questions about ScreenMind's privacy? privacy@shakvaro.com or open an issue at GitHub Issues. ScreenMind is built by Shakvaro (Chattogram, Bangladesh).

Back to ScreenMind