Tried a couple new AI services tonight. Supposedly
Sonauto gives you more control, and lets you name an artist you want the output to mimic. It sort of works, but often sounded garbled, like a backwards played record. Maybe it was just the way I was using it.
Then I stumbled upon
Udio and generated this song:
Canvas ReverieSimilar to Suno, it generates pairs of output in small chunks, and you have to add sections to extend it into a full song. It seems to have more allowance for free use, and the output sounds less compressed than Suno. Generated lyrics can be just as bizarre though.