To access the chat context from within a tool when the tool is defined outside the agent class initialization, follow these steps: 1. Define the Tool with Access to RunContext: Ensure the tool includes the

RunContext

as a parameter. This provides access to the current session and chat context.

from livekit.agents import function_tool, RunContext

@function_tool()
async def my_tool(context: RunContext, other_params: str) -> dict:
    # Access chat context via context.chat_ctx
    chat_context = context.chat_ctx
    # Use the chat context as needed
    return {"result": "success"}

2. Add the Tool to the Agent Dynamically: After defining your agent, add the tool to its list of available tools using the

update_tools()

method.

class MyAgent(Agent):
    async def __init__(self):
        super().__init__(instructions="My agent instructions")
        # Add the tool to the agent's tools
        await self.update_tools([my_tool])

3. Access Chat Context Within the Tool: The

RunContext

object passed to your tool includes the current chat context. You can access it via the

context.chat_ctx

property.

@function_tool()
async def my_tool(context: RunContext, user_id: str) -> dict:
    chat_context = context.chat_ctx
    user_history = chat_context.items
    # Access the conversation history or other chat context details
    return {"user_history": user_history}

Let me know if you need further clarification or a more specific example!