VRChat Voice Input native runtimes

This repository contains the minimal Windows x64 native ASR runtimes used by VRChat Voice Input. Model weights are distributed separately. The application downloads these files from an immutable Hugging Face revision and verifies the size and SHA-256 of every file before installation.

The archive intentionally excludes development tools, test executables, quantizers, Parakeet utilities, SDL, and command-line programs that the application does not call.

Sources

Paraformer resident worker: FunASR commit 5990412a196518d511bff12584417195fb9c952b plus the product worker and English post-processing patch.
SenseVoice resident workers: SenseVoice commit fbf91f3baccebec554dcee68708b0cda61d42805 plus the product worker patch.
Fun-ASR-Nano resident workers: FunASR commit 5990412a196518d511bff12584417195fb9c952b plus the product worker and Vulkan backend patch.
Whisper CPU and Vulkan servers: whisper.cpp v1.9.1 commit f049fff95a089aa9969deb009cdd4892b3e74916.

The FunASR, SenseVoice, and whisper.cpp-derived files are provided under their respective MIT licenses included in each runtime directory. Vulkan-enabled runtimes require the Vulkan loader installed by the target machine's graphics driver; vulkan-1.dll is not redistributed.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support