User Control Flow

UserControlFlowTools enable agents to pause execution and request input from users during conversations.

Prerequisites

Set up a virtual environment with the SDK setup guide, then install:

uv pip install agno openai

Example

The following agent can request user input during conversations:

from agno.agent import Agent
from agno.tools.user_control_flow import UserControlFlowTools

agent = Agent(
    instructions=[
        "You are an interactive assistant that can ask users for input when needed",
        "Use user input requests to gather specific information or clarify requirements",
        "Always explain why you need the user input and how it will be used",
        "Provide clear prompts and instructions for user responses",
    ],
    tools=[UserControlFlowTools()],
)

run_response = agent.run("Help me create a personalized workout plan")

while run_response.is_paused:
    for requirement in run_response.active_requirements:
        if requirement.needs_user_input:
            for field in requirement.user_input_schema:
                if field.value is None:
                    field.value = input(f"Enter {field.name}: ")

    run_response = agent.continue_run(
        run_response=run_response,
    )

print(run_response.content)

The run pauses each time the agent calls get_user_input. Fill the requested fields and call continue_run(run_response=run_response) to resume in memory. To resume by run_id instead, configure a database so the agent can retrieve the paused run. See Dynamic User Input for the full pattern.

Toolkit Params

ParameterTypeDefaultDescription
instructionsOptional[str]NoneCustom instructions for user interaction behavior.
add_instructionsboolTrueWhether to add instructions to the agent.
enable_get_user_inputboolTrueEnable user input request functionality.
allboolFalseEnable all functions.

Toolkit Functions

FunctionDescription
get_user_inputPause agent execution and request input from the user.

Developer Resources