---
title: Unity WebGPU Local Voice LLM
emoji: 📚
colorFrom: blue
colorTo: purple
sdk: docker
pinned: false
license: other
short_description: Voice-in, voice-out LLM, running on Unity and WebGPU
thumbnail: >-
  https://cdn-uploads.huggingface.co/production/uploads/6319b5b0d47392f426ca7447/7u1is766CJHClbvUkZhOm.png
---

**Copyright (c) 2026 David Addis. All Rights Reserved.**

This repository contains compiled web binaries for a local, browser-based voice-to-voice LLM agent using WebGPU. This is for demonstration purposes only. No permission is granted to download, distribute, modify, or reverse-engineer the compiled application or wrapper logic. The application source code is closed and proprietary.

## About This Demo
This interactive web demo runs entirely locally in your browser using WebGPU. No audio or text is sent to the cloud.

## Credits & Open Source Models
While the application wrapper and logic are proprietary, this demo relies on the following open-weights models running locally:
* **Speech-to-Text:** [Whisper](https://github.com/openai/whisper) by OpenAI (MIT License)
* **LLM Engine:** [SmolLM2](https://huggingface.co/HuggingFaceTB/SmolLM2-1.7B-Instruct) by HuggingFaceTB (Apache 2.0 License)
* **Text-to-Speech:** [Supertonic-2](https://huggingface.co/Supertone/supertonic) by Supertone Inc. (OpenRAIL-M / MIT License)

Please see the `THIRD_PARTY_NOTICES.txt` file in this repository for full license texts and copyright acknowledgments.