ContextQMD
Back to Moonshine

CLI Transcriber

examples/windows/cli-transcriber/README.md

0.0.622.9 KB
Original Source

CLI Transcriber

A command-line application for Windows that transcribes speech using the Moonshine C++ API. You'll need Visual Studio 2022 (or later) with C++ development tools installed.

The GitHub release archive is self-contained: it includes the Moonshine library, headers, the English medium streaming model, and a sample WAV file for testing.

Setup

Download and extract windows-cli-transcriber.tar.gz.

If you're working from a git checkout instead, run download-lib.bat to fetch the library bundle separately, then download models with pip install moonshine-voice and python -m moonshine_voice.download --language en.

Building

  1. Open cli-transcriber.sln in Visual Studio
  2. Select Release and x64 (recommended), or Debug and x64 if you need to step through your own code
  3. Build the solution (Build > Build Solution or press F7)

The prebuilt Moonshine libraries are Release-only (/MD). The Debug configuration uses the same runtime library so it can link against them; you still get unoptimized code and debug symbols for the example app itself.

Alternatively, you can build from the command line:

batch
msbuild cli-transcriber.sln /p:Configuration=Release /p:Platform=x64

Running

Transcribe the bundled sample file

batch
x64\Release\cli-transcriber.exe --wav-path beckett.wav

This uses the bundled models/medium-streaming-en model (architecture 5) and should print a transcript of the Beckett quote.

Live microphone transcription

batch
x64\Release\cli-transcriber.exe

This loads the bundled model, listens on the default microphone, and displays transcriptions in real time. Press Ctrl+C to stop.

Custom model path

batch
x64\Release\cli-transcriber.exe --model-path "path\to\model" --model-arch 5

Notes

  • The application uses Windows Audio Session API (WASAPI) to capture audio from the default microphone
  • Audio is automatically resampled to 16kHz mono if needed
  • Make sure you have microphone permissions enabled in Windows settings when using live capture

Adding Moonshine to your own Project

To use Moonshine Voice in an application:

  • Make sure moonshine-voice-windows-x86_64 is downloaded and accessible.
  • Add moonshine-voice-windows-x86_64\include to the include paths.
  • Add moonshine-voice-windows-x86_64\lib to the linker paths.
  • Add all of the libraries in moonshine-voice-windows-x86_64\lib (bin-tokenizer.lib, moonshine-utils.lib, moonshine.lib, onnxruntime.lib, and ort-utils.lib) to be linked.
  • Ensure that onnxruntime.dll from moonshine-voice-windows-x86_64\lib is copied to the same folder as your executable. This example project does that using a custom build step.