# BrowserAI

A JavaScript SDK that runs large language models, speech recognition, text-to-speech and audio separation directly inside the browser using WebGPU, with no server required.

BrowserAI is an SDK for running production LLMs and related AI models directly in the browser. It uses WebGPU acceleration for near-native inference speed, works fully offline after an initial model download, and switches between multiple backend engines, including MLC, Transformers, Flare and Demucs, depending on the task.

Beyond text generation and chat, it supports speech recognition and text-to-speech, structured output generation with JSON schemas, audio source separation to isolate vocals, drums and bass, and a Web Worker mode so inference does not block the page's UI. It ships pre-configured popular models and installs with `npm install @browserai/browserai`.

Website: https://browserai.dev