A reading aid for people with dyslexia. Scan or paste a text, read it in a calmer layout, listen to it sentence by sentence, and ask questions about it. Everything runs on the device: no account, no server, nothing uploaded.
Built with Expo SDK 54 (React Native 0.81, React 19). Runs on iOS, Android and the web.
- Add text: scan with the camera, choose a photo, open a .txt file, paste, or type. On-device OCR uses ML Kit (Android) and Apple Vision (iOS) in native builds, and tesseract.js on the web.
- Reader
- Simplified view: plainer words ("utilize" becomes "use"), filler words like "furthermore" removed, long sentences split. Hold a sentence (or use "Show original wording" in focus mode) to see what it said before. Dates, times and amounts are never altered.
- Dates and times are marked in yellow and amounts in green.
- Reading settings: Atkinson Hyperlegible, Lexend or the system font; text size; line spacing (1.2–2.5×); letter spacing; a colored background tint; high contrast; focus mode (one sentence at a time); a reading ruler (dims all but two lines).
- Listen: tap any sentence to hear the text from there. The sentence being read is highlighted, focus mode follows along, and you can pause, resume, skip and change the voice speed.
- Summary: the most central sentences, the key dates and amounts, and a reading-time estimate. Share it or copy it.
- Ask: type a question to find the sentence that answers it, then read it aloud or copy it. If nothing in the text matches, it tells you so rather than guessing.
- Saved texts: the last 50 texts are kept on the device. Open, rename, pin (up to 10 pinned texts are never dropped), search, delete or clear them.
All reading settings are saved on the device.
Requires Node.js 20.19.4+ and Yarn via Corepack.
corepack yarn install
corepack yarn web # open in the browser
corepack yarn start # dev server for Expo Go / dev builds- Expo Go (quickest): run
corepack yarn startand scan the QR code. Everything works except scanning text from photos, which needs native modules that Expo Go doesn't include. If your phone isn't on the same network, usecorepack yarn start --tunnel. - Development build (full features, including OCR):
iOS builds target iOS 15.5 or later, which ML Kit requires; this is set in
npx expo prebuild corepack yarn android # or: corepack yarn ios (macOS + Xcode)app.json.
- OCR on the web downloads the tesseract.js engine and English language data (about 10 MB) from a CDN the first time you scan. The browser caches them after that, and your images are processed in the browser and never uploaded.
- Text-to-speech uses the browser's Web Speech API, so the available voices depend on the browser and OS.
- Sharing falls back to copying to the clipboard where the browser has no share sheet.
- On a Mac, the web version is the way to run ReadAble. Expo doesn't build native macOS apps.
corepack yarn test # Jest (jest-expo preset)
corepack yarn test:coverage
npx expo install --check # dependency versions match the SDKSee testing-guide.md for a manual test pass and improvements.md for ideas that aren't built yet.
App.js state wiring and screen layout
src/
theme.js light and high-contrast color tokens, font faces
components/ ui.js primitives + one component per card
hooks/
useDocumentProcessor.js active document and saved history
useSpeech.js sentence-by-sentence text-to-speech
useSettings.js persisted reading preferences
utils/
textProcessing.js sentences, dates/amounts, simplify, summary, Q&A
storage.js AsyncStorage: saved texts and settings
ocr.js / ocr.web.js native (ML Kit) and web (tesseract.js) OCR
dialogs.js alerts and confirmations that also work on web
constants/ settings defaults and limits, sample text
Saved texts store only { id, rawText, createdAt } plus an optional custom title and pinned flag the reader set. Everything else (sentences, summary, simplified text) is derived when a text is opened, so improvements to the text processing apply to texts saved earlier too.
- Text processing is rule-based and English-only. Summaries are extractive, and Q&A matches keywords rather than meaning.
- Voices and speech quality depend on the device.
Apache 2.0. See LICENSE.