Input Transformation Pre-Hook

Use a pre-hook to transform a Team’s input before it reaches the LLM.

Code

from typing import Optional

from agno.agent import Agent
from agno.models.openai import OpenAIResponses
from agno.run.team import TeamRunInput
from agno.session.team import TeamSession
from agno.team import Team
from agno.utils.log import log_debug


def transform_input(
    run_input: TeamRunInput,
    session: TeamSession,
    user_id: Optional[str] = None,
    debug_mode: Optional[bool] = None,
) -> None:
    """
    Pre-hook: Rewrite the input to be more relevant to the team's purpose.

    This hook rewrites the input to be more relevant to the team's purpose.
    """
    log_debug(
        f"Transforming input: {run_input.input_content} for user {user_id} and session {session.session_id}"
    )

    # Input transformation agent
    transformer_agent = Agent(
        name="Input Transformer",
        model=OpenAIResponses(id="gpt-5.2"),
        instructions=[
            "You are an input transformation specialist.",
            "Rewrite the user request to be more relevant to the team's purpose.",
            "Use known context engineering standards to rewrite the input.",
            "Keep the input as concise as possible.",
            "The team's purpose is to provide investment guidance and financial planning advice.",
        ],
        debug_mode=debug_mode,
    )

    transformation_result = transformer_agent.run(
        input=f"Transform this user request: '{run_input.input_content}'"
    )

    # Overwrite the input with the transformed input
    run_input.input_content = transformation_result.content
    log_debug(f"Transformed input: {run_input.input_content}")


print("🚀 Input Transformation Pre-Hook Example")
print("=" * 60)

# Create the team members
research_agent = Agent(
    name="Research Analyst",
    model=OpenAIResponses(id="gpt-5.2"),
    role="Expert in market research, data analysis, and competitive intelligence",
)

strategy_agent = Agent(
    name="Strategy Consultant",
    model=OpenAIResponses(id="gpt-5.2"),
    role="Specialist in business strategy, planning, and decision frameworks",
)

financial_agent = Agent(
    name="Financial Advisor",
    model=OpenAIResponses(id="gpt-5.2"),
    role="Expert in financial planning, investment analysis, and risk assessment",
)

# Create a financial advisory team with the pre-hook
team = Team(
    name="Financial Advisory Team",
    model=OpenAIResponses(id="gpt-5.2"),
    members=[research_agent, strategy_agent, financial_agent],
    pre_hooks=[transform_input],
    description="A professional financial advisory team providing investment guidance and financial planning advice.",
    instructions=[
        "Work collaboratively to provide comprehensive financial advice.",
        "Coordinate your expertise to deliver actionable investment guidance.",
        "Always remind users to consult with a licensed financial advisor for personalized advice.",
    ],
    debug_mode=True,
)

team.print_response(
    input="I'm 35 years old and want to start investing for retirement. moderate risk tolerance. retirement savings in IRAs/401(k)s= $100,000. total savings is $200,000. my net worth is $300,000",
    session_id="test_session",
    user_id="test_user",
    stream=True,
)

Usage

1. Set up your virtual environment

   Mac

   uv venv --python 3.12
   source .venv/bin/activate
   

   uv venv --python 3.12
   .venv\Scripts\activate
   
2. Install dependencies

   uv pip install -U agno openai
   
3. Run example

   python input_transformation_pre_hook.py