Agent

The agent workflow itself is inherently sequential -- there is only one "active agent" at a given time controlling the system

If you want parallelization, just put agents as tools to other agents imo

@Logan M thank you for your response. Based on doc wise, if I wanna make parallelization, do you have an example code to follow to understand how it works?

Also, currently agent can't react and execute any task, it's just work as simple chat Q&A type, if I want to do some automation with agents like it to create auto calendar events for me and write email automatically, what way I can do? Any suggestions on these two things