Tools

Tools are the hands of your Agent. They allow the LLM to interact with the outside world, such as searching the web, querying a database, or performing calculations.

Purpose

Use @tool to: - Expose Python functions to the LLM. - Define clear descriptions and type hints so the LLM knows when and how to use them.

Usage

Defining a Tool

Simply decorate a Python function with @tool. Type hints and docstrings are mandatory as they are used to generate the schema for the LLM.

from yosrai import tool

@tool
def calculate_tax(amount: float, rate: float = 0.1) -> float:
    """
    Calculates the tax for a given amount.

    Args:
        amount: The total amount.
        rate: The tax rate (default 0.1).
    """
    return amount * rate

Async Tools

You can also define async tools for non-blocking I/O operations.

import asyncio
from yosrai import tool

@tool
async def fetch_data(url: str) -> str:
    """Fetches data from a URL asynchronously."""
    await asyncio.sleep(1) # Simulate I/O
    return "Data content"

API Reference

yosrai.engine.core.tools.tool(func)

Decorator to mark a function as an Agent Tool. Auto-generates .json_schema for the tool and enforces type validation.