To debug a LiveKit Python agent, you can follow these steps: • Run Agent in Development Mode: Start your agent in

dev

mode for development and debugging. Use the command: `python agent.py dev`This mode connects your agent to LiveKit and makes it accessible from anywhere on the internet while providing debugging capabilities. • Use Built-in Debugging Features: The LiveKit agents framework includes several debugging tools: ◦ Asyncio debugging: Enables asyncio debugging with

loop.set_debug(args.asyncio_debug)

. ◦ Slow callback detection: Configures

loop.slow_callback_duration = 0.1

(100ms) and hooks slow callbacks with

utils.aio.debug.hook_slow_callbacks(2)

. ◦ Exception logging: Automatically logs worker failures with

logger.exception("worker failed")

. • Configure Logging: Set up logging using the framework's configuration: `setup_logging(args.log_level, args.devmode, args.console)`You can also define custom logging in your agent code:

import logging
logger = logging.getLogger("your_agent_name")
logger.setLevel(<http://logging.INFO|logging.INFO>)

• Use Console Mode: For local debugging, run the agent in

console

mode. This keeps the agent local and accessible only within your terminal. • Testing with Playground: Use the Agents playground to test and debug your agent interactively when running in

dev

mode. • Collect Debugging Information: If you need further assistance, gather the following details: ◦ Library version:

pip freeze |grep livekit

◦ Python version:

python --version

◦ Relevant logs with context ◦ Room info (e.g.,

RM_nnnnn

) ◦ Time the issue occurred and timezone ◦ Whether you're using self-hosted or cloud LiveKit Sources: Voice AI quickstart | LiveKit Docs | Voice AI quickstart | LiveKit Docs | livekit-plugins/livekit-plugins-inworld/livekit/plugins/inworld/log.py | Still Have Questions