Text Extract from Photo: OCR

Text Extract from Photo reads printed words out of pictures and hands them back as text you can actually use. Point the camera at a page, a sign or a label, or hand it a stack of photos you already took. Recognition happens on your phone: no account, and no picture is uploaded to us. 🗂️ A whole stack in one run • Pick several photos and they are read in order, one after another • A progress row per photo: waiting, reading, done, or skipped • One unreadable photo never stops the batch — the rest still finish • The results arrive as one combined text to copy, share or export • The summary says how many were read and how many held nothing 🔊 Hear it read back • Play, pause, stop, and a speed dial from slow to brisk • Reading follows the recognised sentences, so you can look away • Good for proofreading a label, or for anyone who prefers to listen • No voice for your language? The app says so and opens the system voice installer instead of pretending to read 📤 Beyond the clipboard • Save as a plain .txt file for any notes or documents app • Save as a PDF you can keep, print or send on • Files go out through the share sheet, so no storage permission is needed 📐 The layout survives Line breaks and paragraphs come out as printed — an address stays an address, opening hours stay a list. Nothing is flattened into one blob. For receipts and two-column pages, switch on column order and each column is read top to bottom before the next begins. 📷 Camera or gallery • Frame the text, tap the shutter, and the words appear in a second • Already have the picture? Import it with the system photo picker • A light toggle for dim rooms, a bag, or a shelf label in the shade 🧩 Blocks, not a wall of text Text arrives split into blocks the way it was laid out. Tap one to pick it — the address, the phone number, the paragraph you wanted. Hold a block to take everything. 🔗 Numbers, emails and links A phone number, email or web address becomes a chip. Tap it and the matching app opens. Nothing opens on its own — a scanned poster can never dial or browse for you. 🗃️ A history you control Every extraction stays on this phone: search the full text, star what you return to, select several to copy, share or delete together, or clear the lot. Kept a photo with an entry? Read it again with a different script when a sign turns out not to be Latin. Turn history off and everything stored goes with it, photos included. 🌏 Latin, Korean, Japanese, Chinese Latin covers English and most European languages and is on by default. Korean, Japanese and Chinese recognisers are included too — pick the one you photograph most. Each script is a separate model, which is why this app is a little larger than a note app; the settings screen says so plainly. ⚙️ Small things that help • Optionally copy the first block as soon as text is read • More than 30 interface languages, switchable in the app • Dark, high-contrast layout, large buttons, a one-handed bottom bar 👋 Meet Quill Quill, the little lens with a text cursor, greets you on first run and explains why the camera is needed. While reading, Quill's eye tracks along the lines. 🔒 Private by design Recognition runs entirely on your device with Google ML Kit's on-device models. Photos and recognised text are not sent to our servers — we do not have any. The only Android permission asked for is the camera, and only when you use it. Importing pictures needs none. ✨ Free, with one optional upgrade Everything above is free, supported by a small banner and an occasional full-screen ad. A one-time purchase removes ads forever; it is not a subscription. Honest about the limits: printed text works well; handwriting, stylised lettering, low light or blurry print may not. Non-Latin PDFs are stored as page images — use .txt export to select the words. 🔎 Premium: drag a box over just the part you need to read.

Fixed the Copy button looking disabled when nothing is selected — it's now clearer that it's still tappable. Recent list, result screen, and area-select screen now load lighter and faster. • Thumbnails in the recent scans list load faster • Result screen preview displays more smoothly • The area-select recognition screen is lighter too Text recognition accuracy stays exactly the same.