SnapRead — Screenshot to Text, Offline OCR

Drag to select any text on screen and copy it instantly. 100% on-device OCR — no uploads, no login, no tracking.

Screenshot to Text — Offline OCR that reads any text on screen, locally Some text just can't be selected. Text baked into a screenshot, a diagram, a scanned PDF, a video frame, a paywalled chart, an error dialog. SnapRead extracts text from any image on screen with one shortcut. Press Ctrl+Shift+O (Cmd+Shift+O on Mac), drag a box around the text, and it's on your clipboard. No file to save, no upload, no waiting on a server. Genuinely private — by design, not by promise Most "free OCR" extensions ship your screen to a remote server. SnapRead can't, because it has no network access at all. The entire OCR engine and English language model are bundled inside the extension and run in WebAssembly on your own machine. The extension requests zero host permissions, which means it is structurally incapable of contacting any website. Open DevTools and watch the Network tab during a capture — it stays empty. How it works Press Ctrl+Shift+O (Cmd+Shift+O on Mac) or click the toolbar icon on any page. The page dims. Drag a rectangle over the text you want. Release. The text appears in a card right where you selected — with a one-click Copy button. Need it again later? Open Recent Reads from the side panel and copy it a second time. Press Esc any time to cancel. New: Recent Reads Copied something, then lost it to your clipboard's next copy? Recent Reads keeps your last captures in a side panel so you can copy them again — no re-capture needed. Right-click the SnapRead toolbar icon and choose "Recent reads", or assign your own keyboard shortcut Keep your last 10, 20, or 50 captures — your choice Copy or delete any entry, or clear everything in one click Stored only on this device, in local browser storage. Nothing is uploaded, because nothing can be Pause it any time, and it never records captures made in Incognito windows Features ⌨️ One keyboard shortcut — capture without leaving the keyboard 🎯 Precise region select — grab exactly the text you need, nothing else 🔒 Fully offline OCR — no network requests, ever 🚫 No account, no telemetry, no data collection 📋 Instant copy — result card with a Copy button, right at your cursor 🕘 Recent Reads — re-copy past captures from a side panel, stored on-device 🖥️ Retina / HiDPI accurate — crops match exactly what you dragged 🌗 Light and dark mode ⚡ Fast — powered by Tesseract and WebAssembly, running natively on your CPU Use it for Copying text out of screenshots, memes, and infographics Grabbing error messages from screenshots without retyping them Pulling text from charts, diagrams, and slide decks Reading text in images on pages that block selection Getting quotes out of scanned documents and non-selectable PDFs Capturing text from video frames and presentations Collecting several snippets in a row, then copying them one by one from Recent Reads Privacy SnapRead collects nothing. No analytics, no crash reports, no identifiers. Captured images are processed in memory and discarded — never saved, never logged, never transmitted. Extracted text is saved to Recent Reads in your browser's local storage on this device only, so you can copy it again; you can pause that, delete individual entries, or clear all of it at any time, and uninstalling SnapRead removes it entirely. Nothing ever leaves your machine. Currently supports: English (Latin script), optimized for clean printed text. More languages coming.