Samuel Colvin

Thanks so much for this. I'll get someone to look at it soon.

`debug.format()` takes the same arguments `debug()` but returns the formatted string instead of printing it.

do you mean the results of a call? This should be possible with raw python, but I need to understand more about what you mean?

Seems a reasonable thing to do, @pedroallenrevez is that what you want? It's worth noting you could do this already with your own logic, potentially using [`functools.cache`](https://docs.python.org/3/library/functools.html#functools.cache).

Thanks everyone for your patience, I've proposed #541 which I hope documents multi-agent applications pretty well. Full support for graphs is coming too, see #528 and #539. I'd love your...

not yet, I want to add multi-modal suport - images, audio and video, but that will come a bit later.

Let's try this in #760.

Many of the models ollama supports seem to also be supported by Groq, see https://ai.pydantic.dev/api/models/groq/. Can you explain why need you need ollama as well as groq?

Okay, happy to support it, especially if we can reuse some of the openai integration. PR welcome

> I'd be happy to give this a crack this weekend unless someone else is already cracking on? @benomahony go for it. I think we need to: * add a...