[sdenton4]

This looks like an excellent step towards being able to apply consistency to generated images across a series.

[basch]

It looks as if it would trivially integrate into Whisk, which already has a similar feature for defining an outputs “subjects” “scene” and “style”.