Image to Text (OCR) — Privacy Policy

Image to Text copies text out of any image or screen region by recognizing it entirely on your machine. Local-first is the whole product, not an afterthought: your images are never uploaded, and the extension makes no network requests at all.

What we collect

Nothing. There is zero telemetry and no network requests whatsoever: no analytics, no usage data, no crash reports, no account. This is enforced in the build, not just promised — a red-line check (scripts/check-network.mjs) fails the release if any outbound-network or remote-code pattern appears in our own shipped code, and the manifest pins connect-src 'self' so pages cannot reach any server.

The recognition engine (Tesseract), its WebAssembly core, and its language data are all bundled inside the extension and loaded from chrome-extension:// URLs. Nothing about your images, the pages you visit, or the text you recognize is collected or sent anywhere.

What the extension stores, and where

• Your preferences (recognition language, "keep line breaks" toggle) are stored locally using the browser's extension storage (storage.sync, key ocr.settings.v1) so they follow your browser profile. They contain no personal data.
• Your interface-language choice (if you pick one instead of following the browser) is stored locally in storage.local under ocr.locale.v1; it never leaves your device.
• A captured screen region, on its way from capture to the recognition page, is held briefly in storage.session (ocr.pendingCapture.v1) and consumed once — it lives only until the browser closes and is never written to disk or synced.
• A small local error log (storage.local, ocr.errlog.v1, last 25 entries) exists only so you can copy diagnostics into a bug report yourself. It is never sent anywhere and contains no image data.
• The images and text you recognize are processed in memory and are never stored, uploaded, or transmitted.

Permissions and why they are requested

The extension installs with no access to any website — no host permissions and no declared content scripts, so it can see no page until you invoke it.

• activeTab, scripting — let you capture a region of the current tab when you click the toolbar icon or press the shortcut. This is a one-time action on that single tab (screenshot + crop), with no permanent access and no host permission.
• storage — stores your preferences and the handoff/error data described above, all locally.

There are no optional permissions and no host permissions, static or optional.

Data sharing

None. No data is collected, so none is shared, sold, or transferred.

Contact

Questions: email [email protected], use the support link in the extension's options page, or use the website listed on the store page. This policy may change as the product evolves; material changes will be noted here with a new date. (This extension is an independent product and is not affiliated with Google or Mozilla.)