Feature requests

It would be very helpful, to extend an uploaded audio file with not just the 32s model, but with the 2:11.

Basically, opening up the manual mode will allow you to select exactly what instruments are used, lead vocal, back up, and so forth. Each user should be able to select these and set them as defaults. Maybe even a few presets as they may have several music types they work on. Each "vocal" should have its own seed number so that a creator can begin to create similar sounds with more consistency.

Please add the ability to upload a melody and have Udio do an arrangement of that exact melody in different styles and genres. Suno has this ability and it is GREAT.

a feature called /describe. The same as in Midjourney, where you can upload an image and it will give you 4 prompt options to recreate a similar imagery. This feature would allow us to upload our music and get a prompt back which would then generate similar to source file outcomes. Benefits: 1) That allows you to understand outputs better 2) prompt more efficiently 3) consistency; if you work on a larger project within your daw

Sometimes I like one verse from one extension and another verse from a different extension made at the same time. I'd like the ability to combine these two verses that originated in different gens (they both have same genre, singer and musical accompaniment). It would also be useful for adding extensions made outside of Udio to tracks. Even better if the feature could be even more powerful, so AI could also be asked to recognise similar / identical parts of the music and replace over at that point automatically. Given the trim feature this could extend use to the user to allow them to effectively cut, of copy and paste tracks. Though a separate interface for this could also be added.

Being able to select a section of track to edit, but instead of changing what is there, you can add extra instruments/vocals over the top. Being able to have this as 2nd editable layer would be useful.

Give us a feature to train/upload our own voicebanks.

To make the vocals so much easier for males and females. You can have a toggle centered at 0. If you shift it 100% to the right, you get female... if you shift 100% to the left, you get male. You also have a (+) button to add more vocals (maybe 10 max, for now) And the toggle is defaulted at the center. That makes 2 vocals if you want to create an easy duet. If you want to add all 10 and mix them up between males and females, that can create an acapella group or opera or theatrical. This would help with the gendered vocals and duets.

For example, it would be cool if we could upload a song (original), and then upload another song (reference), whatever style it is, and convert the original song to the style of the reference song. An example: A 70s song converted to an 80s Bonnie Tyler style, or an 80s track converted to a 2000s style.

I want to be able to set the song name for h next generation before I generate. Renaming the song afterwards results in having to rename up to 6 iterations manually in worst case.