Grab Text from
Anywhere on Your Screen

GrabText is a free Chrome extension that lets you draw a selection box over any part of your screen to instantly copy the text to your clipboard. It first tries to read text directly from the page's HTML (instant). If no text is found — for example on an image or video — it automatically takes a screenshot of your selection and runs OCR (optical character recognition) locally in your browser using Tesseract.js.

Yes. When you draw a selection over an area that doesn't have selectable HTML text — such as an embedded image, a video frame, a canvas element, or a screenshot — GrabText automatically switches to OCR mode. It captures a screenshot of your selected region and runs Tesseract OCR locally to recognise the text, then copies it to your clipboard.

Absolutely nothing is sent to any server. All OCR processing happens locally inside your browser using WebAssembly. No images, text, or usage data ever leaves your device. GrabText does not require an account, an API key, or an internet connection to function.

GrabText supports 11 languages out of the box: English, Hindi, Bengali, Punjabi, Urdu, Spanish, German, Portuguese / Brazilian, Sinhala, Tamil, and Kannada. Before drawing your selection you can pick the OCR language from a dropdown — your choice is remembered for next time.

GrabText is completely free with no usage limits, no watermarks, no premium tier, and no expiry. You can make as many text extractions as you like. There are no hidden fees — the extension is and will always remain free for everyone.

GrabText works on virtually every website including complex apps that use Shadow DOM such as Gmail, YouTube, and Twitter. It is Shadow DOM aware, so it captures text from elements that standard browser selection tools miss. The only exception is the Chrome Web Store and browser internal pages (chrome://) where extensions cannot run by design.

GrabText requests three permissions: activeTab (to inject the selection overlay into the page you're viewing), scripting (to run the selection and text-extraction code), and offscreen (to run Tesseract OCR in a hidden background document without slowing your page). No browsing history, no tabs data, no storage access beyond saving your language preference locally.

After each extraction GrabText shows a result card. The plain text (whitespace collapsed, ready to paste) is auto-copied to your clipboard immediately. The card also shows a formatted text area where line breaks are preserved — you can highlight a portion of it and click "Copy Formatted" to copy just that selection. The card auto-dismisses after 10 seconds or you can close it manually.

Yes. Press Esc at any time while the selection overlay is active to cancel without copying anything.

OCR requires taking a screenshot, cropping it to your selection, and running a machine-learning model (Tesseract) on the pixel data — all locally in your browser. This typically takes 1–3 seconds depending on the size of the selected area and your hardware. DOM extraction reads text directly from the page's HTML tree, which is instant. GrabText always tries DOM first and only falls back to OCR when no selectable text is found.