transcribe.io is a cross-platform mobile application built with .NET MAUI and C#, enabling private, real-time transcription using on-device AI models.

Features

• Record audio directly from your microphone on any mobile device
• Transcribe recorded audio and show transcription result using on-device AI model
• Real-time transcription mode (live display of recognized text)
• Model selection and management (download and switch between models)
• Auto-feature to automatically choose which language is spoken

Tech Stack

• .NET MAUI for cross-platform deployment (iOS, Android)
• C# programming langugage for modern, type-safe application logic
• On-device inference using Whisper.net
• Offering different Whisper variants from Whisper.cpp

Getting Started

1. Clone the repository.
2. Open the solution in your preferred IDE (e.g., JetBrains Rider, Visual Studio, VS Code).
3. Build and run the app.

TODOs

• Implement basic UI with on-device transcription support
• Add support for real-time transcription
• Support downloading and running different Whisper models
• Add iOS deployment support
• Add Android deployment support
• Add support to set transcription language
• Configure CI workflow for basic application testing
• Configure CI/CD for automated binary builds
• Implement unit and e2e tests with automation
• Improve accuracy on real-time transcriptions

Acknowledgements

I want to acknowledge DIPS AS for giving me the opportunity to develop this open-source software.

This project was heavily inspired by MauiWhisper. Real-time transcription was based on the work of whisper_streaming.