Does this skill handle the audio processing loop?

This skill provides the dispatch layer, API routes, and UI indicators. The actual continuous audio processing loop typically runs as a separate background worker process (e.g., LiveKit Agents framework).

Can I create custom AI personalities for my meetings?

Yes, the skill includes a persona system where you can define custom system prompts, specific voices, and choose between proactive or reactive speaking styles.

What is the AI Voice Room skill?

It is a specialized Claude Code capability that allows developers to programmatically dispatch AI participants into LiveKit video rooms to listen, process, and speak in real-time.

What are the prerequisites for installation?

You need a Next.js app with the App Router, along with the video-room, ai-core, and transcription skills already installed.

Which transcription and voice services are supported?

By default, it uses Deepgram for Speech-to-Text (STT) and is configured to work with common Text-to-Speech (TTS) providers through the ai-core and transcription dependencies.

AI Voice Room Agent

Name: AI Voice Room Agent
Author: mattwoodco

bymattwoodco

•

コラボレーションツール

Deploys real-time AI voice participants into LiveKit video rooms to assist with meetings, strategy sessions, and interviews.

The ai-voice-room skill enables developers to integrate interactive AI agents directly into live video environments using LiveKit. These agents function as full participants that listen via Deepgram transcription, process context through customizable LLM personas, and respond using high-quality text-to-speech. Whether you need a passive note-taker to summarize key decisions or a proactive strategist to challenge assumptions, this skill provides the dispatch layer, persona management, and state tracking necessary to build sophisticated voice-first AI applications within a Next.js environment.

主な機能

01Automated participant token management and agent dispatching API

02Real-time voice interaction via LiveKit and Deepgram STT/TTS

03Extensible architecture for custom system prompts and specialized voices

04Visual state tracking for idle, listening, thinking, and speaking modes

053 GitHub stars

06Predefined personas including Note-taker, Strategist, Interviewer, and Tutor

ユースケース

01Interactive mock interviews or language tutoring sessions using specialized personas

02Live brainstorming support where an AI strategist provides constructive feedback and challenges assumptions

03Automated meeting minutes and action item tracking with a Note-taker agent

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add mattwoodco/skills ai-voice-room

For use in Claude.ai and ChatGPT

Download Skill