Hey everyone! 👋 I'm just starting my DevOps journey and recently tried using AgentGPT for some automation tasks. I was really excited about it at first, but I hit a wall when I realized I couldn't run it locally for my projects that handle sensitive data.
That's why I switched to CrewAI. The main draw for me was being able to run everything on my own machine. As a beginner, it felt safer and more aligned with the self-hosted tools I'm learning about (like Docker and local K8s clusters).
Here's a basic `docker-compose.yml` I used to get a CrewAI example running locally:
```yaml
version: '3.8'
services:
crewai-setup:
image: python:3.11-slim
working_dir: /app
volumes:
- ./:/app
command: >
sh -c "pip install crewai &&
python my_crew_script.py"
```
It's not much, but it worked! Having everything local means I can tinker without worrying about external APIs or data leaks. Has anyone else made a similar switch? I'd love to hear about your experiences or any beginner tips for setting up local AI agent workflows.