Building a Simple Multi-Agent System with OpenAI’s Swarm

OpenAI’s Swarm provides a framework for orchestrating multi-agent systems. Below is an example of how you can use the framework to create agents that perform specific tasks.

from swarm import Swarm, Agent

client = Swarm()

def transfer_to_agent_b():
    return agent_b


agent_a = Agent(
    name="Agent A",
    instructions="You are a helpful agent.",
    functions=[transfer_to_agent_b],
)

agent_b = Agent(
    name="Agent B",
    instructions="Only speak in Haikus.",
)

response = client.run(
    agent=agent_a,
    messages=[{"role": "user", "content": "I want to talk to agent B."}],
)

print(response.messages[-1]["content"])

Explanation:

• Swarm Client: Initializes the Swarm system.
• Agent A: Acts as a helpful agent, but transfers control to Agent B.
• Agent B: A creative agent that only responds in Haiku format.
• Execution: The user talks to Agent A, which passes the task to Agent B.

This structure allows for multi-agent collaboration, where agents can shift tasks among each other dynamically. For more details, visit the Swarm GitHub page.