I've been working on this for AGES but here is a very, very early beta of what will evolve over the next few months.

There are two new models:

- Maestro-3.0-L

-Maestro-3.0-prompted

Maestro-3.0-L is designed for general-purpose generation (with pretty much non-existent prompting) and Maestro-3.0-prompted is trained a bit more on prompts to a point where you get something you want more than with no prompt, but definitely not to a point where it really understands what you mean.

Also the "guide ai" button is in the bottom left if you can't find it.