Basic Usage

This guide covers how to create simple agents using the Kibo SDK.

Defining an Agent

The AgentConfig acts as the DNA of your agent.

from kibo_core import AgentConfig

agent_def = AgentConfig(
    name="MathTutor",
    description="You are a patient math tutor.",
    instructions="Explain the solution step-by-step.",
    agent="agno",
    model="gpt-4o-mini"
)

Creating the Instance

Use the create_agent factory. This function detects the requested engine (agno, langchain, etc.) and returns the appropriate adapter.

from kibo_core import create_agent

agent = create_agent(agent_def, api_key="sk-...")

Running an Agent

All Kibo agents expose a synchronous run method and an asynchronous run_async method.

Synchronous

result = agent.run("What is the square root of 144?")
print(result.output_data)  # "The square root of 144 is 12."
print(result.metadata)     # Contains usage info, adapter name, etc.

Asynchronous

import asyncio

async def main():
    result = agent.run_async("What is pi?") # Returns a KiboFuture
    # To wait for the result:
    final = result.result() # Blocking call on the future
    # OR in a truly async loop (with asyncio.to_thread wrapping)
    # See Distributed Execution section for best practices.