READ BY YOUR MACHINE, NOT A CLOUD
UtilitiesOCR — text out of images and PDFs, on your machine
Scanned pages, photographed documents, screenshots of text — Tesseract reads the words out, entirely on your device, and hands you the text.
Runs locally in your browser. Nothing is uploaded.
Text trapped in an image is text you cannot search, quote or edit — and the usual escape route uploads the whole document to someone's OCR service. This tool takes the other route: Tesseract, the industry's open-source workhorse, running entirely on your device. The scan is read where it already is, and the text comes back to you.
It handles both shapes the job arrives in: single images (the photographed whiteboard, the screenshot, the scanned receipt) and multi-page PDFs (the whole scanned chapter, the archive document). Point it at the file; get the words.
The engine downloads once on first use, then stays cached — after that, OCR works offline, which is both a convenience and a statement: the reading happens on your hardware, at your speed, with no per-page metering and no account.
The extracted text is ordinary text — feed it to Notes, search it, quote it, run it through the transcript tools, or typeset it into a fresh document. Free, private, and immediate.
What you can do
The parts people actually reach for.
Tesseract on-device
The open-source OCR workhorse, running on your machine.
Images and multi-page PDFs
Screenshots and whole scanned chapters alike.
Offline after first load
The engine caches once; after that, no network needed.
Text you can use
Searchable, quotable, editable — ordinary text out.
Feeds the workspace
Send extracted text to notes, quizzes and the transcript tools.
No upload, no account
Documents stay on your device while they are read.
How it works
Three steps, no detours.
Open the image or PDF
It loads locally — the page or scan stays on your machine.
Run the recognition
Tesseract reads the text on-device, from its cached engine.
Take the text
Ordinary, usable text — for notes, search or anywhere else.
Questions, answered
The ones people ask first.
Which languages does OCR support?
Tesseract's language set — the common languages are covered; pick what the document is in.
How big is the engine download?
Roughly 75 or 500 MB depending on the model — once, then cached for offline use.
Are my documents uploaded?
No. Recognition runs entirely on your device.
More in Drazil
One workspace, 44 tools — all free, all local.
Runs locally in your browser. Nothing is uploaded. · or browse all 44 tools