Difference between using Agents and the Hamilton library in WrenAI

[BraianTuck]

I’m exploring WrenAI to build intelligent data pipelines, and I’ve noticed that there seem to be two different approaches: using Agents or using the Hamilton library.

I understand that:
Agents provide flexibility and allow integrating LLMs to make dynamic decisions during pipeline execution.
Hamilton focuses more on defining data transformations and dependencies in a clear, function-based way.
However, I’m not entirely sure when it’s best to choose one approach over the other.

My question:
What are the main differences between using Agents and using Hamilton within WrenAI?
Are there specific use cases where one is more suitable than the other?
How do these two approaches compare in terms of scalability, maintainability, and ease of debugging?

Thanks in advance for any clarification!