generate the missing part.
play one part — a loop, a bassline, a hummed melody — and the part that isn't there gets written to fit it, at your tempo.
one gpu, shared with everything else running. if it's slow, that's why.
1 · your part
hum, play, or drop a file. anything past 120 seconds gets trimmed — the model is trained on 30, and it audibly thins out over long takes.
2 · what's missing
3 · which model
picking a model wakes it. a model nobody has touched for two days takes about 90 seconds to come back — that happens while you record, not after you press generate. otherwise switching is instant.
4 · the sound
genre
character
placement
these are not free text. the model was trained on a closed set of words — anything outside it is a phrase it has never seen, and it renders worse. this list is that set, and it changes with the stem: drums and bass do not share character words.
the words are v0.2/v0.3's vocabulary. v0.1.5 learned a different, older phrasing, so writing your own caption suits the two newer models and pushes v0.1.5 off what it knows. leave the genre alone and each model gets its own.
sending:
5 · generate
steps
seed
measured on a 30-second part: 64 steps holds the gpu 5.7 s, 32 holds it 3.4 s, 24 holds it 2.8 s. 64 is the standard everything here is evaluated at, and the difference is audible — it is the default for that reason, not because it is slower.
the same prompt swings a long way between seeds — measured up to 0.4 on byte-identical input. if a take is bad, take a new seed before you rewrite anything.
the take