ocr-it
Chrome extension: pin a screen region once, then hotkey your way through a paginated document. OCR runs 100% offline via bundled Tesseract.
What is it?
What it is
OCR It is a cross-browser (Chrome 116+ and Firefox 140+) Manifest V3 extension that lets a user pin a rectangular region on any paginated document viewer (scanned books, slide decks, PDFs, embedded readers) and then repeatedly capture and OCR that region with a hotkey. It supports single captures, an unattended auto-run that captures and turns pages until the document ends, and exports the resulting transcript as text or a .txt file. OCR runs entirely on-device via a bundled Tesseract build.
Why it exists
Many documents on the web — scanned books, slide viewers, image-based PDFs, readers that block text selection — make their content unreadable to copy/paste, to LLMs, or to search. OCR It exists to liberate that trapped text. It is built around an offline-first, privacy-respecting philosophy: no API key, no network calls, no images leaving the machine, and no host permissions requested at install time. The auto-run and reliable end-detection exist so a user can walk away from a multi-hundred-page job and come back to a clean transcript, including ones long enough to feed into an LLM for summarisation or question-answering.
Who should use it
Users who need to extract text from non-selectable paginated documents (scanned books, PDFs in viewers, slide decks) and value offline, private processing without API keys.
Who should avoid it
Users working only with selectable text, those needing to process over 300 pages in a single automated run, or anyone unable to install browser extensions.
Stars over time
+42 stars in the last 2 days
How it works
A quick walkthrough in plain English
How ocr-it works
Step 1 of 3
You interact with it
Open ocr-it, send a request, or connect it to your stack.
Features
Advantages
- Privacy‑first offline OCR, no keys or data leaving the machine
- One‑time setup and hotkey convenience for fast batch capture
- Auto‑run reduces manual effort for long documents
- Robust page‑turn support for embedded viewers and shadow DOM
- Duplicate detection prevents endless loops
- Editable per‑page results and easy export
- Multi‑language support (English/Portuguese/Spanish + add‑ons)
- Reliable unattended runs with automatic end detection
- No install‑time permissions – only opt‑in when needed
- Open‑source MIT licence, easy to extend
Disadvantages
- Requires manual region selection and occasional re‑pick if zoom/resize changes
- Auto‑advance breaks if the window size or zoom changes mid‑run
- Chrome PDF viewer cannot be turned automatically (must page manually)
- Local PDFs need extra “Allow access to file URLs” in Chrome or a permission grant in Firefox
- Accuracy depends on source quality; low‑res scans or handwriting need cleanup
- Captures are limited to the visible viewport
- Rate‑limited to a couple of screenshots per second; fast hotkey mashing only queues
- Adding new languages needs vendoring and editing `src/shared.js`
- Firefox lacks a built‑in shortcut editor; users must navigate `about:addons`
- Page cap of 300 may truncate very long books
- No real‑time OCR preview – you edit text after capture
- Duplicate detection may stop early on legitimate repeated content
- No support for internal browser pages (`chrome://`, `about:`)
Installation
cloud
Chrome 116+: https://chromewebstore.google.com/detail/ocr-it/pekiamimgjbjgpggeilimnaplklceabh | Firefox 140+: https://addons.mozilla.org/en-US/firefox/addon/ocr-it/
native
Chrome (116+): Open chrome://extensions, enable Developer mode, click Load unpacked and select build/chrome (or the repo root directly since it is a valid extension as checked in). Firefox (140+): Open about:debugging#/runtime/this-firefox, click Load Temporary Add-on and select build/firefox/manifest.json. Alternatively, npm run start:firefox launches a scratch profile with the extension installed and auto-reloads on edits. Build from source: git clone https://github.com/thiagotigaz/ocr-it.git && cd ocr-it && npm run build
FAQ
How do I install OCR It?
Install from the Chrome Web Store (Chrome 116+) or Firefox Add-ons (Firefox 140+). After installing on Chrome, check chrome://extensions/shortcuts to ensure the hotkey is set, since Chrome may leave it blank if another extension claims it. Firefox users should go to about:addons → gear icon → Manage Extension Shortcuts.
What hotkeys does OCR It use?
Three main hotkeys: Option+Shift+S (capture the region once), Option+Shift+A (start/stop an automatic run), and Option+Shift+R (draw or redraw the capture region). On macOS, Option is the same as Alt.
How do I set up automatic page turning?
Enable 'Turn the page automatically after capture' in settings, then use 'Pick control' to click the viewer's next-page button, or press a key (default ArrowRight) which dispatches a keyboard event into the frame owning the middle of your capture region. Use 'Test now' to verify it works before starting a long run.
Can I add more languages for OCR?
Yes. Three languages ship by default (English, Portuguese, Spanish). To add more, run 'npm run vendor -- fra deu jpn' to pull Tesseract language models into vendor/lang/, then add the language codes to the LANGUAGES array in src/shared.js. Each language adds roughly 0.7-3 MB to the extension.
Why should I leave the Layout setting on Auto?
Auto runs Tesseract's layout analysis which correctly handles multi-column layouts and returns text in reading order. Other modes like 'Single block' or 'Single column' can silently fail on multi-column pages, producing interleaved text with high confidence but poor accuracy - measured at 52.9% character error rate versus 0.0% on Auto.
Does OCR It work offline?
Yes, 100% offline. OCR runs locally with a bundled Tesseract build - no API key, no network, and no images leave your machine. The extension makes no outbound requests at all.
Featured in Videos
YouTube tutorials and walkthroughs for ocr-it
Alternatives
Similar projects ranked by category, topics, and text overlap.
