To help you reduce conversation history to the last 2–4 turns while still capturing the important context, here are a few friendly approaches you might find useful: • Truncate Chat Context: You can use LiveKit’s

truncate()

method to keep just the last 2–4 message exchanges. This way, you retain the most recent conversations along with the initial system message that holds your instructions. • Dynamic Truncation: Another great option is to use the

before_llm_cb

callback in the

VoicePipelineAgent

, which lets you dynamically limit the context right before each LLM call. • Manual Summarization: You might also consider creating a summary of earlier interactions with a custom method like

UserData.summarize()

, then adding it as a system message when transferring agents or resetting context. Since LiveKit doesn’t have built-in summarization APIs, you’ll need to handle summarization yourself. If you’d like, I’d be more than happy to share some detailed code examples for any of these methods—just let me know! Sources: livekit-plugins/livekit-plugins-aws/livekit/plugins/aws/experimental/realti... | agents/src/llm/chat_context.ts | VoicePipelineAgent | LiveKit Docs | VoicePipelineAgent | LiveKit Docs | complex-agents/medical_office_triage/README.md