Practical overview
When it fits
Use Whisper.net when a .NET project needs speech-to-text functionality through Whisper models.
Who it is for
- .NET developers adding speech transcription to an application.
- Developers creating a transcription utility that uses Whisper models.
Documented uses
- Add speech-to-text transcription to a .NET application.
- Build a transcription utility using Whisper models through .NET bindings.
Requirements & limits
- This is a library for developers to integrate, rather than a ready-to-use transcription application.
- Its documented stack is .NET bindings for OpenAI Whisper through whisper.cpp.
- The supplied project details do not specify setup steps, runtime requirements, supported operating systems or deployment modes.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Install Whisper.net with the CPU runtime
Add Whisper.net’s CPU inference runtime to a .NET project.
Before you start
- A .NET project that can accept NuGet package references.
- For the default CPU runtime: Windows 11 or Windows Server 2022 or newer and the Visual C++ Redistributable (Visual Studio 2022 or newer); Linux with libstdc++6 and glibc 2.31; macOS prerequisites are listed as TBD.
- On x86/x64, a CPU supporting AVX, AVX2, FMA and F16C; otherwise use the NoAvx runtime.
Steps
Add a reference to the Whisper.net package, version 1.9.1, in the project file.
Add a reference to Whisper.net.Runtime, version 1.9.1, for the CPU runtime.
For a CPU without the listed AVX-family instruction support, select Whisper.net.Runtime.NoAvx instead of the default runtime.
Expected outcome
The project references Whisper.net’s core package and a CPU runtime; the README describes the default runtime as CPU inference.
Guide limitations
- This guide covers package setup only; it does not document model acquisition or execution steps.
- The README lists macOS prerequisites as TBD.
- The standard x86/x64 runtime requires AVX, AVX2, FMA and F16C; NoAvx is identified for CPUs lacking AVX support.
Transcribe a WAV stream with Whisper.net
Process a WAV file stream with a Whisper model and print timestamped text results.
Before you start
- A .NET project with Whisper.net and a runtime package installed.
- A model file named ggml-base.bin and a WAV file whose path is available as wavFileName.
- Runtime prerequisites for the selected platform and hardware.
Steps
Create a WhisperFactory from the model file path ggml-base.bin.
Create a processor with the language set to auto, then build it.
Open the WAV file as a read stream.
Enumerate the processor’s ProcessAsync results and print each result’s start time, end time and text.
Expected outcome
The example writes each processed result’s start and end times and text to the console.
Guide limitations
- The example does not explain how to obtain or select a model, or how to prepare the WAV file.
- The example does not document error handling, output-file writing or accuracy expectations.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 13 October 2026
- Original website title
- GitHub - sandrohanea/whisper.net: Whisper.net. Speech to text made simple using Whisper Models · GitHub · Official homepage title · retained in its source language