Local speech-to-text transcription using Whisper models with ONNX Runtime. Zero-configuration, automatic GPU acceleration, and lazy model downloading.
Get Started
$ dotnet add package LMSupply.TranscriberReadme
LMSupply.Transcriber
Local speech-to-text transcription using Whisper models.
Features
- Zero-config: Models download automatically from HuggingFace
- GPU Acceleration: CUDA, DirectML (Windows), CoreML (macOS)
- MIT Licensed: OpenAI Whisper models
- 99+ Languages: Auto-detection and multilingual support
Quick Start
using LMSupply.Transcriber;
// Load the default model
await using var transcriber = await LocalTranscriber.LoadAsync("default");
// Transcribe audio
var result = await transcriber.TranscribeAsync("audio.wav");
Console.WriteLine(result.Text);
Console.WriteLine($"Language: {result.Language}");
Console.WriteLine($"Duration: {result.DurationSeconds}s");
Available Models
| Alias | Model | Size | WER | Description |
|---|---|---|---|---|
fast | Whisper Tiny | ~150MB | 7.6% | Ultra-fast |
default | Whisper Base | ~290MB | 5.0% | Balanced |
quality | Whisper Small | ~970MB | 3.4% | Higher accuracy |
medium | Whisper Medium | ~3GB | 2.9% | High quality |
large | Whisper Large V3 | ~6GB | 2.5% | Highest accuracy |
english | Whisper Base.en | ~290MB | 4.3% | English optimized |
GPU Acceleration
# NVIDIA GPU
dotnet add package Microsoft.ML.OnnxRuntime.Gpu
# Windows (AMD/Intel/NVIDIA)
dotnet add package Microsoft.ML.OnnxRuntime.DirectML