What is the right method to get json data from the Agent in funtion_tool @function_tool async def update_answers(self,answers:dict[str,Any]): self.session.userdata.data_collection.update_answers(answers) help me with this

Thanks for reaching out, Rajadurai. I'm an AI assistant and I'm here to help. Give me a moment while I research this topic.

how to give decription for answers parameter?

You're asking how to add a description for the

answers

parameter in your

@function_tool

. Two clear options: 1) Docstring (simple)

@function_tool()
async def update_answers(self, context: RunContext, answers: dict[str, Any]):
    """
    Update collected answers in session userdata.

    Args:
        answers: A dict mapping question IDs to the user's answers.
                 Example: {"q1": "foo", "q2": "bar"}.
    """
    # ...

2) Typed annotation with

Annotated

+

Field

(recommended for LLM/tool schema)

from typing import Annotated
from pydantic import Field

@function_tool()
async def update_answers(
    self,
    context: RunContext,
    answers: Annotated[dict[str, Any], Field(description="Dict mapping question IDs to answers, e.g. {'q1':'foo', 'q2':'bar'}")]
):
    # ...

Use the docstring for human readers; use

Annotated

+

Field

when you want the description included in the tool/function schema sent to the LLM. Which approach do you prefer or are you targeting LLM function-calling behavior? Sources: Building Multi-Agent Architectures with LiveKit Agents | How to test your agent using another agent | Building a Moviefone-style Theater Assistant | LiveKit docs | Tool definition and use | LiveKit docs | Azure OpenAI LLM plugin guide | LiveKit docs

livekit.agents._exceptions.APIStatusError: Error code: 400 - {'error': {'message': "Invalid schema for function 'update_answers': In context=('properties', 'answers'), 'additionalProperties' is required to be supplied and to be false.", 'type': 'invalid_request_error', 'param': 'tools[1].function.parameters', 'code': 'invalid_function_parameters'}} (status_code=400, request_id=147b26d6-725b-419e-b678-88f678963bd3, body={'message': "Invalid schema for function 'update_answers': In context=('properties', 'answers'), 'additionalProperties' is required to be supplied and to be false.", 'type': 'invalid_request_error', 'param': 'tools[1].function.parameters', 'code': 'invalid_function_parameters'}, retryable=True)

Thanks — I’m responding to the schema validation error you posted for

update_answers

(the error says the

answers

property must include

additionalProperties: false

). Two safe fixes: 1) Quick fix — supply a raw JSON schema with

additionalProperties: False

for `answers`:

raw_schema = {
  "name": "update_answers",
  "description": "Update answers with question IDs and values",
  "parameters": {
    "type": "object",
    "properties": {
      "answers": {
        "type": "object",
        "description": "Dict mapping question IDs to answer values",
        "additionalProperties": False
      }
    },
    "required": ["answers"]
  }
}

2) Recommended — let LiveKit auto-generate a correct schema by using type hints / Pydantic instead of a manual schema:

from livekit.agents import function_tool
from typing import Any

@function_tool()
async def update_answers(self, answers: dict[str, Any]) -> str:
    """Update collected answers (mapping question IDs to values)."""
    # implementation...
    return "Answers updated"

Which approach are you using now — raw schema or @function_tool/type hints? Share that and any current parameter definition and I’ll give the exact code to fix it. Sources: Tool definition and use | LiveKit docs