​
Overview

Humanlayer integrates seamlessly with ControlFlow to add human oversight to your AI applications. This guide shows you how to use Humanlayer with ControlFlow.

​
Installation

Install the required packages:

pip install humanlayer controlflow python-dotenv

​
Basic Example

Let’s create a simple math agent that performs basic arithmetic operations with human oversight.

from dotenv import load_dotenv
from humanlayer import HumanLayer
import controlflow as cf

load_dotenv()

hl = HumanLayer(
    # run_id is optional - it can be used to identify the agent in approval history
    run_id="controlflow-math",
)

def add(x: int, y: int) -> int:
    """Add two numbers together."""
    return x + y

@hl.require_approval()  # require user approval before this function can be called
def multiply(x: int, y: int) -> int:
    """multiply two numbers"""
    return x * y

@cf.task(tools=[add, multiply])
def do_math() -> int:
    """multiply 2 and 5, then add 32 to the result"""

if __name__ == "__main__":
    result = do_math()
    print("\n\n---------- RESULT ----------\n\n")
    print(result)

​
How it Works

  1. We create a HumanLayer instance with an optional run_id to identify this agent in the approval history
  2. Define two math functions: add and multiply
  3. Use @hl.require_approval() to require human approval before the multiply function can be called
  4. Create a ControlFlow task that uses these functions as tools
  5. The agent will request approval before executing the multiply function

​
Running the Example

  1. Set up your environment variables in .env
  2. Run the example:
python main.py

When the agent attempts to use the multiply function, it will pause and request human approval through your configured contact channel (Slack, Email, etc).

​
Next Steps

For more framework examples, see https://github.com/humanlayer/humanlayer/tree/main/examples