Winner - Google Chrome Built-in AI Challenge 2025 (Most Helpful Application)
AAC Board AI is an Augmentative and Alternative Communication board that helps people who cannot rely on speech communicate more easily and naturally. It is enhanced with Built-in AI for on-device proofreading, rewriting, and translation — keeping interactions private, fast, and reliable offline.
Try the live demo at aacboard.app. Requires Chrome or Edge with Built-in AI enabled.
- Grammar Correction – Turns telegraphic text into clear sentences (Proofreader API).
- Tone Adjustment – Rewrites messages in direct, professional, or friendly tones (Rewriter API).
- Translation – Translates messages and boards between languages (Translator API).
- Text to Speech – Reads messages aloud (Web Speech API).
- Offline Ready – Installs as a standalone app with automatic updates (PWA).
- Open Board Format – Imports
.obfand.obzfiles (examples).
Grammar correction and tone adjustment require experimental browser flags.
Chrome:
chrome://flags/#proofreader-api-for-gemini-nano
chrome://flags/#rewriter-api-for-gemini-nano
Edge:
edge://flags/#edge-llm-rewriter-api-for-phi-mini
Requires Node.js 24+.
git clone https://github.com/shayc/aac-board-ai.git
cd aac-board-ai
npm install && npm run devnpm run dev # Start dev server
npm run lint # Lint
npm test # Run tests
npm run build # Typecheck + production build- UI: React 19, TypeScript, Material UI, React Router
- AI: Gemini Nano (Chrome) / Phi 4 Mini (Edge)
- Speech: Web Speech API
- Storage: IndexedDB
- Data: open-board-format (OBF/OBZ file parsing)
- Tooling: Vite, Vitest, Playwright
See full architecture details in docs/architecture.md.
Load a board by passing an OBF or OBZ file URL as the board query parameter:
https://aacboard.app/?board=https://example.com/board.obz
The app fetches, imports, and navigates to the board automatically.
While I'm thrilled if you find this code useful, this repository serves as my personal workspace and I am not accepting pull requests at this time. Feel free to fork the project and adapt it to your own needs under the MIT license.
MIT © Shay Cojocaru
