Research Symfony AI Components

Random thoughts:

- How do they handle platform with all the metadata about providers and models?
- Do they have some concept of "Operation Types", some way of mentioning the skills of
- How do they handle multiple providers and models coming together with a consistent approach to logging (Such as cost metadata?)
- How do they handle tool logging, and tools within the code not just external tools..
- Do they have some consistent manner of handling permissions in the tools?
- The Agents, what theoretical approach to agents do they take (Swarm? Orchestrator? All?).
- Do they have my concept of agents creating a "workflow" so that php handles the orchestration of tasks not AI handling the orchestration of tasks. For example "go through these 100 piece of content and check for PII"

(Reading through their docs and answering it)

I think it could provide some nice interfaces and models for us to use.

So from conversations with the team and doing some research, our next steps will be:

1. Locate improvements to the AI Platform.
2. Look into how integration of the AI Platform would work.
3. At a later stage when its stable, take a decision to make the move for AI Platform and possibly AI Store.

Do we need a meta issue to track these tasks?

Thanks, that's a great idea - I have added one here 🌱 [Meta] Use Symfony AI Active . Please comment what you see is missing or add general discussion points there.

This issue where basically just to make sure I had time to research it before a meeting we had with them, thus it being very sparse.

Automatically closed - issue fixed for 2 weeks with no activity.

unassigning