Skip to main content
Demonstrates memory growth performance for a reasoning-enabled team.
The source calls a streaming team four times without consuming any returned async generator. Those calls never execute, so the evaluation does not build the intended history or memory. Update the saved file before running.
team_response_with_memory_and_reasoning.py

Run the Example

1

Set up your virtual environment

2

Install dependencies

3

Export your OpenAI API key

4

Run PgVector

5

Execute all four team runs

Set stream=False and stream_events=False on the Team, then replace each _ = team.arun( in run_team_for_user with _ = await team.arun(.
6

Run the example

Save the code above as team_response_with_memory_and_reasoning.py, then run:
Full source: cookbook/09_evals/performance/team_response_with_memory_and_reasoning.py