Extensions / OCR Scanner
OCR Scanner privacy policy
Last updated: November 19, 2026
OCR Scanner turns an image or a visible page area into editable text. This policy distinguishes processing inside the extension from the separate websites you may choose to visit.
What the extension collects
The extension has no account system, cloud OCR service, advertising tracker or third-party analytics inside the extension. Its OCR code does not upload your images, recognized text or edits to the developer or an OCR provider. It does not read your browsing history or build an OCR history.
You choose the source: an image file, an image you paste, or a selected area of the current web page. For area capture, Chrome temporarily captures the visible tab; the extension crops that image locally. The page selector handles the selection coordinates, not the page’s text content.
What stays in your browser
Image bytes, the local image preview, recognized text and your edits are held in the live result document and local processing contexts. They are not written to extension storage as a saved history. Closing the result window discards the working result; reopening it does not restore that draft.
Your chosen OCR language and the review prompt state (up to five local dates of successful use, whether you answered, dismissal count and a snooze date) are kept in the extension’s own IndexedDB database in your browser. A successful-use date is recorded after a successful Copy of an OCR result. This information is used locally to decide when to show the review prompt; it is not extension analytics sent to us. These settings do not use Chrome Sync. If an earlier build kept them in chrome.storage.local, the extension moves them into that database once at startup and deletes the old copy.
chrome.storage.session holds routing information such as result-window, source-tab, source-window and source-document IDs and, while an area selection is open, the state of that selection (tab, window and document IDs, a one-time token, status flags and an expiry time), so a capture can finish if Chrome restarts the extension’s background worker. It does not hold the image or OCR text. Removing the extension removes its extension storage, including that database, through Chrome; it does not erase text you have copied or pasted elsewhere.
Copy writes the current text to your system clipboard only when you click it. Pasting an image is handled only in response to your paste action in the source chooser. The extension does not read your clipboard in the background. Clipboard contents can remain available to your operating system and other apps after the result window closes. We do not promise to erase operating-system memory, caches, backups or clipboard history.
Network requests
The OCR engine, WebAssembly files and five language models (English, German, French, Spanish and Russian) are included in the extension package. The current OCR configuration loads those local resources, disables its model cache and has no remote OCR API or cloud fallback. These five OCR languages are separate from the languages used for the store listing.
Opening an external website is a separate activity from OCR. The website host receives ordinary web requests when you visit it. The website section below describes the service pages separately; it is not a claim that every part of the product or browser is offline.
Pages the extension opens and website services
The extension’s lifecycle code is designed to open a one-time welcome page after installation, not after an update. Chrome can open the uninstall page after removal. Choosing one to three stars in the review prompt can open the separate feedback page; higher ratings lead to Chrome Web Store reviews. The extension does not attach your image, recognized text, edits or source-page URL to these links.
The three service pages are hosted on GitHub Pages at welcome, uninstall and feedback. GitHub receives ordinary website requests when you visit a page it hosts; this is separate from local OCR processing.
The service pages use Umami website analytics, using the publisher’s shared website identifier, to count visits to welcome, uninstall and feedback pages. When the analytics script runs, the browser sends visit information to Umami, which can include the page URL and title, referrer, browser, operating system, device and screen information. Hosting and analytics providers also receive the network information needed to serve requests. Website analytics is separate from the extension, which has no analytics and does not send OCR content. The privacy policy page has no Umami script.
Both feedback forms are optional. If a form is not connected, clicking Send feedback sends nothing and keeps your input on the page. When Google Forms is connected, the uninstall form sends a chosen reason and an optional comment; the feedback form sends its single field, “What should be better?”. Neither form requests an email address. Clicking Send feedback sends the entered fields through Google Forms, without automatically adding OCR content or a source-page URL. Please do not include sensitive information in your message. A no-cors send request cannot itself confirm that Google received a response; the page keeps your text rather than claiming confirmed receipt.
GitHub, Umami and Google process website and form requests under their own policies. These website services are separate from local image recognition in the extension.
Permissions, explained
- activeTab gives temporary access to the current tab after you invoke the extension, for the area-selection and visible-tab capture workflow. It is not continuous access to every website.
- scripting lets the extension insert its temporary area selector into the permitted current page after you choose Select area. The selector is not a permanent script on every site and is not used to read the page’s text.
- storage keeps the session routing IDs and the open-selection state described above and lets the extension move settings from an earlier build out of
chrome.storage.localonce. The OCR-language preference and review-prompt state are kept in the extension’s IndexedDB database. Neither is used to save your images or recognized text.
The manifest requests no host permissions and no clipboardRead, clipboardWrite, tabs or offscreen permissions. Some pages restrict selection or capture; the extension does not obtain broader site access to bypass those restrictions.
Changes
When this policy changes, its date and text will be updated at the policy address on pinwisp.com.
Contact
For questions about OCR Scanner or this policy, open an issue in the service-page repository. Issues are public; do not include private images, recognized text, personal information or other sensitive data in an issue.
OCR Scanner is an independent project and is not affiliated with or endorsed by Google LLC. Chrome and Google are trademarks of Google LLC. Tesseract is an independent open-source OCR project; its included licenses and notices remain in the extension package.