I use Synthesizer V Studio 2 Pro (which Herb referred to as Synth V above) from Dreamtonics (https://dreamtonics.com/synthesizerv/).

- the vocal tracks it generates can be indistinguishable from actual recorded vocal tracks
- they are NOT using stolen vocals like Suno and other AI services, instead, they are more like the BIAB model of paying artists to perform and then licensing those performances
- they offer tons of voices but it can get rather expensive!
- to generate a vocal track you can play it on your MIDI keyboard, "draw" the notes directly using the MIDI editor, import a MIDI melody or import a recorded vocal track with melody, timing and lyrics detected automatically (but imperfectly, so you'll need to do a fair amount of cleanup)
- I am very happy with all of their voices that are American English but not at all interested in those with an Asian accent as they don't work for my style pf music
- for a male American English type of voice I like Liam a lot
- for a female American English type of voice I like Etta a lot
- male voices I have are Liam, Hayden, Kevin and Asterian
- female voices I have are Etta, Felicia, Natalie, and Solaria (I also have Mai who has an accent but she was free so...)

I have thus far only used this tool for generating harmony parts to my vocals but I could see this being super helpful for generating high-quality demos and even for releasing as finished works if you don't plan to sing your songs yourself.