If you want to create a voice service that will run in every room:

meshagent service create --name "voice" --image docker.io/meshagent/sample-voice:0.0.31 --env="MESHAGENT_PORT=8090" --port "num=8090 type=meshagent.callable liveness=/ path=/agent participant_name=voice"

If you want to test the voice with a single room:

meshagent service test --room my-room --name "voice" --image docker.io/meshagent/sample-voice:0.0.31 --env="MESHAGENT_PORT=8090" --port "num=8090 type=meshagent.callable liveness=/ path=/agent participant_name=voice"

Here is the voice service code if you want to create your own app:

from meshagent.livekit.agents.voice import VoiceBot
from meshagent.api.services import ServiceHost

import asyncio
import os

service = ServiceHost()

@service.path("/agent")
class SampleVoiceAgent(VoiceBot):
    def __init__(self):
        super().__init__(
            name="meshagent.livekit.agent",
            title="sample voice agent",
            description="sample agent that will respond to a query via voice",
            labels=[ "voice" ],
            auto_greet_prompt="hello",
            rules=[
                "You are a helpful assistant communicating through voice."
            ]
        )

asyncio.run(service.run())