01 / OVERVIEW
Overview
FunASR.NET brings model loading, audio features, neural-network operators, VAD, and Mandarin synthesis into a familiar .NET stack. CLI tools, a web service, and browser demos offer a path from transcribing one recording to experimenting with interactive speech applications.
02 / CAPABILITIES
Key capabilities
01
Mandarin ASR, VAD segmentation, and WebSocket transcription
02
Chinese speech synthesis and browser demos
03
CPU SIMD and an explicitly enabled Vulkan hybrid backend
03 / POSSIBILITIES
Ideas to explore
- Explore Mandarin transcription and speech output in .NET apps
- Experiment with browser microphones, WebSockets, and VAD segmentation
- Work with PCM, G.711, and resampling in telephone-audio experiments
04 / GET STARTED
Getting started
Before you begin
- Install the matching .NET 10 SDK specified in the README
- Prepare and verify compatible model weights separately
- The optional Vulkan backend requires a compatible system driver
Get the project
git clone https://github.com/donet-org-cn/FunAsrNet.gitNext steps
- Clone the source, build the solution, and run the checks that do not require models.
- Prepare the ASR / VAD or Mandarin TTS weights using the model documentation.
- Choose the CLI or web-service example and validate it with your own audio.
05 / SCOPE & STATUS
Current boundaries
An independent community implementation, not an official FunASR SDK. Compatible model weights must be prepared separately. No NuGet package has been published.
This introduction is based on the public README. Consult the repository for current versions, licenses, and supported features.
Updated 2026-10-27