Run StableDAW on your PC with 1 click (No GPU needed) A full AI music studio powered by Stable Audio 3: generate tracks from text, edit them on a timeline, master them, even DJ them --- 100% local. Created by GANTASMO No GPU required. All OS supported: Windows, Linux,... and Mac.show more

1-click install on pinokio

One generation form, four modes: 1. text-to-audio 2. audio-to-audio (hum a melody, get a track back) 3. inpainting (repaint any region of a waveform) 4. continuation 5. Chimera: stack several clips and it blends + beat-aligns them into one brand new generation.

It's not just a generator -- there's a real DAW here. Multi-track timeline with trim & fade handles, a 24-effect mix chain (mastering, reverb, lo-fi, LUFS…), and a step sequencer + piano roll whose patterns feed right back into the AI.

Then it goes full party mode. A two-deck DJ console with beatmatch sync, hotcues, FX rack and Automix, loading tracks straight from your generation library. You can DJ a set of music that didn't exist an hour ago. Plus a 3D audio-reactive VJ engine for the visuals.

You can even teach the model your own sound -- LoRA training is built into the UI (8 adapter types, live telemetry). And the LEARN tab renders your library as a 3D genealogy galaxy: every remix, inpaint and blend traces back to its sources.

It defaults to the Small model, which runs on CPU. But you can switch to the medium model if you have NVIDIA GPU (~8GB VRAM).

Original project github

Full tutorial here to learn more:

@gantasmo

@gantasmo Oooo 👀

@gantasmo The original project was renamed to “theDAW” to avoid confusion around whether or not it’s a Stability AI product, perhaps that ship has sailed here. Glad to see it supported in Pinokio!

@gantasmo aha i see, i was also confused why it used multiple names

@gantasmo this is so sick

@gantasmo Yo @cocktailpeanut , starskreamexe here, cocreator of theDAW. Just wanted to let you know this is a deprecated archived version, the new version is It’s mucheasier to install and has a lot more features. Thanks for the post @gantasmo @DanielJTrujillo

@gantasmo omg i was JUST looking at this earlier but was too intimidated to try and install it O_O

@gantasmo ,

@gantasmo Is anything capable of outputting midi arrangement for the composition? e.g. If I already have human vocals and want a higher quality, non-AI sounding track to go along with it, generating the midi and then assigning my own sound design would create a high quality result quickly.

@gantasmo omg this video gives me so much 2012 Dubstep flashbacks

@gantasmo @AethrMusik check it out

@gantasmo How is the quality compared to Ace Step 1.5?

@gantasmo ace step for songs with lyrics stable audio for songs without lyrics (really good at it)

@gantasmo Lol did one today seeing the most wanted list, planned further testing before publishing. You beat me to it ! 🤣

@gantasmo haha we need to think of an effective way for builders to coordinate so we don't build redundant stuff

@gantasmo Well I am learning it's not lost ! Will work on others. I have an ace step Demon going also so will experience the publishing process with this one instead😁

@gantasmo just make sure to write a nice intro post when you publish so it gets exposure. a lot of users seem to install straight from the official updates posts from the home page

This is a really cool implementation of stable audio 3. Built in LORA is a great idea. Saves me a trip through Colab. Tagging all the clips takes me the most time but it’s nice to have all the features in one place. It’d be cool if this was in a plug-in so I could integrate it with the daw I use.

@gantasmo No GPU needed" for full audio gen means CPU diffusion — that's the catch. What's the actual sec-per-track for a 30s clip on a normal laptop CPU vs offloading to the 4090? Stable Audio's the heavy step; the timeline/DJ wrapper is cheap. Curious where the real floor sits.

@gantasmo this is similar to djstudio? i want to automix multiple songs and render in the fly like djstudio