Hey there! If you want to create a simple

llm_node

hook that prints out the LLM response in LiveKit, you can do this by overriding the

llm_node

method in your custom Agent class. This way, you get to peek at and handle the LLM's output before it moves along the pipeline. Here’s a friendly example to get you started:

import logging
from pathlib import Path
from dotenv import load_dotenv
from livekit.agents import JobContext, WorkerOptions, cli
from livekit.agents.voice import Agent, AgentSession
from livekit.plugins import deepgram, openai, silero

logger = logging.getLogger("llm-response-logger")
logger.setLevel(<http://logging.INFO|logging.INFO>)

load_dotenv()

class PrintingAgent(Agent):
    def __init__(self):
        super().__init__(
            instructions="""
                You are a helpful agent.
            """,
            stt=deepgram.STT(),
            llm=openai.LLM(),
            tts=openai.TTS(),
            vad=silero.VAD.load()
        )
    
    async def on_enter(self):
        self.session.generate_reply()
    
    async def llm_node(
        self, chat_ctx, tools, model_settings=None
    ):
        async def process_stream():
            async with self.llm.chat(chat_ctx=chat_ctx, tools=tools, tool_choice=None) as stream:
                async for chunk in stream:
                    if chunk is None:
                        continue
                        
                    # Extract content from the chunk
                    content = getattr(chunk.delta, 'content', None) if hasattr(chunk, 'delta') else str(chunk)
                    if content:
                        <http://logger.info|logger.info>(f"LLM Response: {content}")
                    
                    yield chunk

        return process_stream()

async def entrypoint(ctx: JobContext):
    session = AgentSession()

    await session.start(
        agent=PrintingAgent(),
        room=ctx.room
    )

if __name__ == "__main__":
    cli.run_app(WorkerOptions(entrypoint_fnc=entrypoint))

Quick rundown: • The

PrintingAgent

class takes charge by overriding the

llm_node

method. • Inside, the

process_stream

async generator goes through each chunk of the LLM’s output. • We grab the content from each chunk and log it so you can see what the LLM is saying. • Then, we yield those chunks so the rest of the pipeline keeps running smoothly. With this setup, you’ll see the LLM’s response printed out in the console before it gets passed to the TTS component. If you want to tweak anything or have questions, just give me a shout—I’m here to help!