Hey Product Hunt 👋 I'm Sam, the maker of Jarvis. I wanted an assistant I could actually talk to, that would then go and do the work on my own computer instead of just chatting. So I built Jarvis: you speak, it answers out loud in about a second, and when the job needs your computer (code, files, research, browsing) it tells you it's on it and hands the task to the AI agent you already use: Claude Code, Codex, Antigravity or Hermes. When the agent finishes, Jarvis reads you the answer. What's inside: 🎙️ Realtime voice with barge-in: interrupt it anytime 🤖 Your agents, on your computer, never ours 🃏 Live answer cards: markets, weather, flights, recipes, charts, images, 3D 🙂 A photorealistic live avatar you can talk to face to face ⌚ iPhone, Apple Watch, Android, desktop (Mac, Windows, Linux) and web Two ways to run it: 1. Open source (MIT): clone github.com/samukbg/Jarvis-Live, bring your own Gemini key, and self-host the client and server. The AVTR-1 live-avatar engine is included. 2. Jarvis Cloud: install the free desktop app, link it to Jarvis Cloud, and talk to your own computer from your phone or watch from anywhere, with no port forwarding. I'd love your feedback: what would you hand to Jarvis first?

Jarvis Live

Open-source voice AI that talks back and runs your agents

Open-source voice AI that talks back and runs your agents