You're asking how to upgrade a Python LiveKit Agent from v1.2 to v1.3 — here’s a concise upgrade path. 1. Upgrade the package • With pip:

pip install --upgrade "livekit-agents[silero,turn-detector]~=1.3"

(include optional extras

silero

and

turn-detector

if you use VAD or turn detection). 1. Verify environment & keys • Confirm Python version and that env vars are set:

LIVEKIT_API_KEY

,

LIVEKIT_API_SECRET

,

LIVEKIT_URL

, and any model keys (e.g.,

OPENAI_API_KEY

). • Reload env if needed (e.g.,

lk app env -w

). 1. Update code if you use new features • Agents 1.3 adds Agent Observability and improved turn detection. If you want multilingual turn detection, swap to `MultilingualModel`:

from livekit.plugins.turn_detector.multilingual import MultilingualModel

session = AgentSession(
  stt="assemblyai/universal-streaming:en",
  llm="openai/gpt-4.1-mini",
  tts="cartesia/sonic-3:...",
  vad=silero.VAD.load(),
  turn_detection=MultilingualModel(),
)

1. Enable Observability (optional) • To use transcripts/traces/logs/audio in LiveKit Cloud, enable Agent observability in your project’s Data and privacy settings. Observability requires SDK v1.3+. 1. Test and deploy • Test locally (Agent Playground or your frontend). • Deploy with the CLI:

lk agent deploy

This performs a rolling deployment. If you want, I can: • Check your current

requirements.txt

or

pyproject.toml

for steps to update, • Show exact code changes for your AgentSession, or • Walk through enabling observability in your project settings. Which would you like? Sources: Agent insights in LiveKit Cloud | LiveKit docs | Voice AI quickstart | LiveKit docs | Improved End-of-Turn Model Cuts Voice AI Interruptions 39% | Deploying to production | LiveKit Docs | Agents v0.x migration guide - Python | LiveKit docs