Back to catalogue
Project
Learning Assistant AI
Mobile prototype that accepts text and images through Gemini and reads responses aloud.
How it’s builtHow it’s built
Learning Assistant AI · Expo mobile study assistant using Gemini for text, vision and voice
I built an Expo React Native study assistant with three tabs: a Gemini 2.5 Flash chat, a visual tab that sends camera or gallery photos of notes to Gemini vision, and spoken answers generated by Gemini's TTS model with the Aoede voice. Responses are cleaned of Markdown before speech, and a custom math helper renders LaTeX as Unicode and 2D fractions.
- app tabs (home, chat, visual)
- 3
- Gemini models used
- 2
- lines of TS/TSX
- ~3,000
Technology and what it does here
- Expo 54 / React Native 0.81
- Cross-platform mobile app with expo-router tabs
- Gemini 2.5 Flash (REST)
- AskGeminiText for chat history and askGeminiVision for image questions
- Gemini TTS
- SynthesizeSpeech returns PCM audio with the Aoede prebuilt voice
- expo-audio
- UseVoice hook plays synthesized speech and tracks speaking state
- expo-camera / expo-image-picker
- Capture or pick photos of notes for the visual tab
- TypeScript
- Lib/math.ts LaTeX-to-Unicode linearizer and display-math tokenizer
Implementation details
- tts.ts parses the sample rate from the returned mime type (e.g. audio/L16;rate=24000) to build playable audio.
- lib/math.ts converts x^2, H_2O, \times and \frac into Unicode inline and splits display formulas for 2D layout, written without React imports so it is unit-testable.
Technologies
- Expo
- Gemini
- Speech
