Privacy Policy — OCR It

Last updated 24 August 2026 · OCR It · source on GitHub

OCR It does not collect, transmit, or sell any data.

The extension makes no network requests. It contains no analytics, no telemetry, no crash reporting, and no remote code. Text recognition runs entirely on your computer using a Tesseract build bundled inside the extension package.

What the extension handles

To do its job, OCR It processes three kinds of information, all only on your own machine:

DataWhyWhere it goes
A screenshot of the visible tab To crop the region you selected Held in memory, cropped, then discarded
The recognised text and a thumbnail of each crop So you can review, edit and export your transcript Local browser storage on your computer
Your settings and the saved region To remember your setup between sessions Local browser storage on your computer

Full-size crops are deleted as soon as a page has been read successfully. Only a small thumbnail and the text remain, so you can check that the region lined up.

No browsing history is kept. A captured page is stored with its text, its thumbnail, its size and the time it was taken — not the address or the title of the page it came from. Nothing the extension stores says which site you were on.

Your control over it

Everything the extension stores is local and yours. Clear in the popup deletes all captured pages, their text and their thumbnails. Removing the extension deletes everything it ever stored.

Nothing is backed up, synced, or shared. There is no account and no server.

Permissions

PermissionWhy it is needed
activeTabTo screenshot the tab you are reading, only after you invoke the extension by hotkey or from its popup
Optional site accessRequested only when you press Allow in the popup, and only for the site you are on. Needed for an automatic run that continues across page loads, or a reader inside a cross-origin frame
scriptingTo draw the region picker and turn the page in the viewer you are reading
storage, unlimitedStorageTo keep your transcript and settings on your computer; a long book can exceed the default quota
offscreenTo run the OCR engine, which needs a canvas and a worker that a service worker cannot provide
tabsTo identify the tab a capture belongs to, and to refuse capturing on browser-internal pages
alarmsTo restart an interrupted automatic run

The extension requests no site access when you install it. It reads a page only when you press its hotkey or a button in its popup, and any durable site access is granted by you, per site, from that popup.

Contact

Questions or concerns: open an issue at github.com/thiagotigaz/ocr-it/issues.