Pashto Voice Recorder - Speech Dataset App
A focused Android app for building high-quality Pashto speech datasets with streamlined recording, review, and session management features. Built with Jetpack Compose and Kotlin.
Technologies Used
App Screenshots
Swipe to view more screenshots
📱 Overview
Pashto Voice Recorder is a focused Android application designed for building high-quality Pashto speech datasets. Built with modern Android development practices using Jetpack Compose, the app streamlines the process of recording scripted prompts, reviewing takes, and organizing recording sessions into a dataset-friendly structure.
✨ Key Features
🎯 Core Functionality
- Pashto Dataset Capture - Streamlined workflow for recording scripted Pashto sentences
- Session Management - Organize recordings by session with CSV file integration
- Simple Recording UI - One-tap recording with runtime permission handling
- Review and Re-take - Navigate to any sentence index for re-recording
- Session-centric Navigation - Route-based navigation using sessionId/csvFileId model
🎨 User Experience
- Material Design 3 - Modern, clean interface following Google’s guidelines
- Theme Support - Light/Dark/System theme options with ThemeMode
- Multilingual Support - English and Pashto language switching
- Responsive Design - Optimized for various screen sizes
🛠️ Technical Stack
Modern Android Architecture:
- Jetpack Compose - Declarative UI framework
- Koin DI - Lightweight dependency injection
- Room Database - Local persistence for structured data
- DataStore Preferences - Local preferences management
- Navigation Compose - Type-safe navigation
- Accompanist Permissions - Runtime permission handling
- Material 3 - Google’s latest design system
- OpenCSV - CSV file parsing for dataset management
- AndroidX DocumentFile - Storage Access Framework integration
🌟 Project Highlights
Technical Achievements
- Focused Dataset Collection - Streamlined workflow for Pashto speech dataset creation
- Modern Android Development practices with Jetpack Compose and Material 3
- Reactive State Management using collectAsStateWithLifecycle and lifecycleScope
- Theme and Language Support with proper locale handling across API levels
- Session-based Architecture with CSV integration for structured data collection
Dataset Impact
- Pashto Language Support - Specialized for collecting Pashto speech datasets
- Open Source Contribution to language learning and AI training initiatives
- Cultural Preservation - Supporting Pashto language digital resources
- Research Enablement - Providing tools for linguistic research and AI development
🚀 Key Technical Features
- Recording Workflow - Start sessions from Home screen and record prompted sentences
- Review System - Navigate to any sentence index for re-recording takes
- Session Navigation - Route-based navigation using sessionId/csvFileId model
- Permission Handling - Runtime audio recording permissions via Accompanist
- Local Storage - DataStore for preferences and Room for structured data
- File Management - Android SAF DocumentFile for persisted tree URIs
- CSV Integration - OpenCSV for parsing dataset prompt files
📊 Project Stats
- Framework: Jetpack Compose + Kotlin
- Architecture: Session-based with Koin DI + Room + DataStore
- Language Support: Kotlin 2.0.21
- SDK Requirements: minSdk 24, targetSdk 36
- Dependency Injection: Koin
- Size: 3.39 MB
- Status: Active Development
- Languages: English, Pashto, Urdu, Dari
🔐 License & Source Code
This is a proprietary project. The source code is not publicly available to protect the intellectual property and business logic.
However, the app demonstrates professional Android development practices and can be showcased as part of a technical portfolio.
📞 Interested in This Project?
If you’re interested in discussing:
- Custom Android app development
- Kotlin and Jetpack Compose expertise
- Speech recognition and audio processing apps
- AI and machine learning mobile applications
- Technical consulting for similar projects
Get in touch - I’d love to hear from you!
A modern Android application focused on Pashto speech dataset collection, built with cutting-edge Android development practices and specialized for linguistic research and AI training initiatives.