sauravpanda/BrowserAI
A browser-based runtime for executing LLMs (llama, deepseek-distill, kokoro) entirely client-side using WebGPU.

BrowserAI is a TypeScript SDK that enables running large language models directly in the browser without server infrastructure. It leverages WebGPU for hardware-accelerated inference, supporting multiple engine backends including MLC, Transformers, Flare, and Demucs. The project includes pre-optimized models for chat, voice interaction, and text-to-speech, enabling fully private and offline-capable AI experiences entirely within the browser environment.