AI voice notes app for iOS, Android & Web
Voxa
Role: Product engineer — design, mobile, AI & backend · Period: 2024 · Client: Own product · Platforms: iOS, Android, Web
Speak, and AI turns your voice into a clean note — title, summary, task list, reminders with exact times and tags. One React Native codebase for iOS, Android and the web, with a live demo you can try right here.
Voice is the fastest way to capture a thought, but a raw recording is hard to act on. Voxa listens, transcribes in real time and lets AI do the organizing: "Tomorrow at 10 call with the client, need to prepare the estimate" becomes a task list and a reminder for exactly 10:00 tomorrow.
It works in English and Russian, keeps notes on the device (offline-first) and never breaks: if the AI is unavailable, a built-in rule-based parser takes over. The whole stack runs on free tiers — built-in speech recognition, Google Gemini and a Netlify Function that keeps the API key off the client.
Key features
- Live speech-to-text in English and Russian
- AI structuring: title, summary, tasks, reminders and tags
- Relative dates resolved to exact times ("tomorrow at 10" → 10:00 tomorrow)
- Tick off tasks, search across notes, view the original transcript
- Offline-first storage with an automatic no-AI fallback
- "Try an example" and typed input for browsers without voice support
What I did
- Designed and built the app end-to-end with React Native, Expo Router and TypeScript
- Implemented cross-platform speech recognition for iOS, Android and the web
- Built a serverless AI endpoint with model fallback, so a busy or rate-limited model never breaks the app
- Wrote a bilingual rule-based parser as an offline fallback
- Shipped the web build with a live, embeddable demo
Results
- Works on iOS, Android and the web from one codebase
- $0 running cost on free tiers
- AI response in a few seconds
Tech stack
- React Native
- Expo
- TypeScript
- Expo Router
- Gemini API
- Netlify Functions
- Speech recognition
- AsyncStorage