Show HN: Ved AI Voice Assistant
TL;DR
Ved is an open-source project for a fully offline, self-hosted AI voice assistant, pitched as a home Jarvis-style system. The developer targets everyday consumer hardware, mentioning a 12 GB VRAM GPU setup that should handle multimodal work without crushing the main machine. Planned next steps include a server-client setup, cheap mic/speaker endpoints around Raspberry Pi-class hardware, and secure remote access beyond the local network.
Nauti's Take
Treat Ved as a lab project for now, not a finished home agent. If you want to test local voice AI, measure latency, speech recognition, multi-room audio routing, and GPU load before counting privacy and offline operation as real wins.
The source base is thin, with solid demos and installation reports still missing.
Briefingshow
Ved points at a real demand: smart-home voice assistants are convenient, but usually cloud-bound and data-hungry. A local assistant that runs on realistic hardware would appeal to tinkerers, privacy-conscious users, and home-server setups. The hard part will not be the pitch, but reliable wake word handling, audio routing, latency, permissions, and long-term maintenance.
Sources