>>713862560Out of the box? For the most part, no. Not unless the dataset it was trained on includes that specific character. LORAs typically solve that.
For animations there's image-to-video as a stopgap, so you could take a preexisting image or screenshot of a character and animate that.