SPCHR (Speech Transcription Tool)

A Windows desktop application that provides real-time speech-to-text transcription using either Azure Speech Services or OpenAI's Whisper model locally.

Features

• Real-time speech-to-text transcription
• Automatic fallback to local Whisper model if Azure credentials aren't available
• Global hotkey support (Ctrl+Alt+L) to start/stop recording
• Automatic text pasting to active window
• Support for both cloud (Azure) and local (Whisper) transcription

Prerequisites

• Windows OS
• .NET 6.0 or later
• Visual Studio 2019 or later (for development)
• Azure Speech Services subscription (optional)

Installation

1. Clone the repository
2. Build the solution using Visual Studio
3. Configure settings in appsettings.json if you want to use Azure Speech Services:

   {
     "AzureSpeech": {
       "SubscriptionKey": "your-key-here",
       "Region": "your-region-here"
     }
   }

Usage

1. Launch the application
2. Use the global hotkey (Ctrl+Alt+L) to start/stop recording
3. Speak clearly into your microphone
4. Transcribed text will automatically be pasted into the active window

Technical Details

• Uses EchoSharp library for real-time streaming transcription
• Uses Whisper.net for transcription
• Implements OpenAI's Whisper model locally via Whisper.net
• Uses NAudio for audio capture

Dependencies

• EchoSharp
• Whisper.net
• Whisper.net.Runtime
• NAudio
• Microsoft.CognitiveServices.Speech

Development

The application uses a modular architecture with:

• Real-time audio capture and buffering
• Background processing for continuous transcription
• Proper resource management and cleanup
• Error handling and fallback mechanisms

License

MIT License

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.