[UPDATE] 8 AI Models - I Have a Problem

[Innermost]

I added 7 more AI models to OBSIDIAN Neural. Nobody asked. I did it anyway.
You can now assign a different "brain" to each of your 8 tracks: piano, vocals, EDM leads, trap beats, 808s, lofi stems, wavy motifs, and one that generates something vaguely musical if you squint. Each model is "specialized", which means it fails in its own unique and personal way.
The GPU network routes your request to whichever provider node is awake. If none are, it falls back to Stable Audio via fal.ai. If that also fails, I recommend a piano.
Is this too many AI models for a plugin that takes 30 seconds to generate a loop? Probably. Am I stopping? No.
It's open source. You can self-host the whole thing and pay exactly nothing, which is what I would do if I were you. There are also subscriptions if you enjoy giving money to a guy who spent his week adding 7 unnecessary AI models.
obsidian-neural.com