Local Speech Transcriber

Local Speech Transcriber is a minimalist voice typing application that uses OpenAI’s Whisper locally.

How It Works

This WPF application captures audio from your microphone and uses a local Whisper-based STT server to transcribe speech, automatically typing the result into any active window.

Development Status

The application is functional and designed with extensibility in mind. Since it already meets my personal usage needs, further development will slow down. Below is a list of potential features I may implement in the future:

• Cross-platform support via a non-WPF presentation layer using MAUI Blazor Hybrid
• Integration of a faster STT solution
• UI options for configuring transcription settings (model, language, etc.)
• Taskbar/tray icon showing the recording state.

Usage

1. Clone this repository. This will also clone Whisper Streaming and SimulStreaming.
2. Verify that Whisper Streaming runs successfully (needs Python and other dependencies).
3. Update appsettings.json if needed. Refer to WhisperSettings for details on the available configuration options.
4. Build and run LocalSpeechTranscriber.sln.
5. Start transcription using the global hotkey (default: Ctrl+Alt+Shift+D) or by clicking the "Start Recording" button.