Hey everyone! I'm just starting to explore CrewAI for a simple automation task. I built a crew that reads product feedback from a CSV, summarizes it with an LLM, and then generates a report.
It works, but it's... slow. Like, 30-40 seconds for just 10 rows of feedback. I'm using the default `gpt-4` model via OpenRouter.
Here's my main agent setup:
```python
from crewai import Agent, Task, Crew, Process
analyst = Agent(
role="Feedback Analyst",
goal="Summarize user feedback concisely",
backstory="You analyze user comments.",
verbose=True,
allow_delegation=False,
llm=llm
)
```
My tasks are sequential: read, analyze, write. Is this expected speed for a small crew? Should I be using a faster LLM or is there a config setting I'm missing?
Maybe I'm doing something wrong because I'm used to a single Lambda call finishing in a few seconds. Is the overhead from CrewAI's coordination just naturally high? Any tips to speed things up would be awesome 😅.