Voice Transcriber

Native Android app for voice transcription serving as a "semantic notebook for conversations."

Architecture

• UI: Jetpack Compose + Material 3 (slate dark theme)
• State: ViewModel + StateFlow
• DI: Hilt
• Persistence: Room Database
• Transcription: Android SpeechRecognizer API (planned)
• Background: Foreground Service with microphone type

Build

export ANDROID_HOME=~/Android/Sdk
export ANDROID_SDK_ROOT=~/Android/Sdk
./gradlew assembleDebug

Requires full JDK (not JRE-headless) for AGP 8.2+ with JDK 21.

Project Structure

app/src/main/java/com/voicetranscriber/
├── App.kt                    # Hilt application class
├── MainActivity.kt           # Entry point, Compose UI
├── ui/theme/                 # Material 3 theme
├── ui/screens/               # Screen composables
├── service/                  # Foreground transcription service
├── data/database/            # Room database, DAOs, entities
├── data/repository/          # Repository layer
└── data/di/                  # Hilt DI modules

Data Model

• RecordingSession: id, title, start/end time, notes
• TranscriptSegment: id, sessionId, text, speakerId, subjectId, timestamp, confidence
• Speaker: id, name, colorHex
• Subject: id, name, parentId (hierarchical)

Permissions

• RECORD_AUDIO
• FOREGROUND_SERVICE
• FOREGROUND_SERVICE_MICROPHONE
• POST_NOTIFICATIONS