Ved AI Voice Assistant

Name: Ved AI Voice Assistant
Availability: InStock
Author: Amigo112

by Amigo112·Jul 9, 2026·1 point·0 comments

Visit Project View on HN

AI Analysis

●MidShip It

LangGraph and Ollama stack when Open WebUI already exists.

Strengths

•Interruptible voice pipeline with mid-sentence barge-in support.
•Human-in-the-loop approval for filesystem and code execution tools.

Weaknesses

•Standard LangGraph plus Ollama stack, no novel architecture.
•Zero stars on GitHub, very early with aspirational roadmap.

Post Description

I built Ved because I wanted a fully offline, self hosted Jarvis for my house, that runs comfortably on everyday consumer hardware (specs 12GB VRAM GPU), without ruining system performance(multi modal).

Where I plan to take this next:

1)Server-client Architecture: I want to shift the heavy load off of ur main device onto a secondary central server where the heavy LLMS and lower models work without ruining ur personal machine.

2)Smart Home: I want to make small cheap mic/speaker endpoints (using small ups like Raspberry) allowing you to communicate with Ved without needing another appliance.

3)Remote Access: I am checking out ways to securely make Ved communicate with outside local networks for free.

I would love to hear from anyone who has worked on a multi room audio setup. What hardware or protocols you found had the lowest latency.

Please try it out and let me know if you have any feedback or suggestions :)