Agents
Agent Teams [Deprecated]
Agent Teams was an initial implementation of our multi-agent architecture (2023-2025) that uses a transfer/handoff mechanism. After 2 years of experimentation, we’ve learned that this mechanism is not scalable and is not the best way to build multi-agent systems.
With our learning over 2 years, we released a new multi-agent reasoning architecture in 2025, please use the new Teams architecture instead.
We can combine multiple Agents to form a team and tackle tasks as a cohesive unit. Here’s a simple example that converts an agent into a team to write an article about the top stories on hackernews.
hackernews_team.py
Run the script to see the output.
How to build Agent Teams
- Add a
name
androle
parameter to the member Agents. - Create a Team Leader that can delegate tasks to team-members.
- Use your Agent team just like you would use a regular Agent.