Show HN: Open-source voice mode and talking avatar for React, in the browser

Category: library

Tags: react, voice, avatar, browser, speech, 3d, llm

Score: 7.3/10 (Innovation: 7, Technical: 8, Documentation: 8, Utility: 6)

react-ai-voice-avatar is a React library that adds full-duplex voice interaction and optional 3D lip-synced avatars to web apps. It runs speech recognition, synthesis, and even LLM inference entirely in the browser or connects to your own providers, eliminating per-minute fees and server infrastructure. The project stands out for its headless hook design, streaming sentence-by-sentence speech, interruption handling, and a 60 FPS 52-blendshape avatar, all with minimal bundle impact.

Target audience: frontend developers, full-stack developers, AI application developers

Repository: https://github.com/927tanmay/react-ai-voice-avatar · TypeScript · MIT · 20 stars

View on Hacker News