Text-to-audio — prompt in, audio out. Audio-to-audio — restyle a source clip with a prompt + init_noise_level. Inpainting / continuation — regenerate a time window inside a clip, or extend the tail …

Browser-based AI audio DAW for Stable Audio 3 with text-to-audio, inpainting, LoRA training, FFmpeg effects, waveform editing, sequencer, piano roll, and persistent library. - gantasmo/StableDAW

Text-to-audio — prompt in, audio out. Audio-to-audio — restyle a source clip with a prompt + init_noise_level. Inpainting / continuation — regenerate a time window inside a clip, or extend the tail …

Full-stack DAW built on Stable Audio 3. FastAPI async-job backend with React 19 / Vite 6 / Tailwind 4 frontend. CREATE/EDIT/TRAIN/LIBRARY tabs, multi-track waveform ...

May 10, 2026 · Goal: Add human-readable guide text above each UI section (paragraph-style) and hover tooltips on individual controls with descriptions and suggested usage.

504 lines (412 loc) · 19.4 KB main StableDAW / stable_audio_3 / training / utils.py Top Code Blame 504 lines (412 loc) · 19.4 KB Raw 1 2 3 4 5 6 7 8 9 10 11 12 13 ...

320 lines (237 loc) · 14.8 KB main StableDAW / docs / guides / ui-controls-guide.txt Top Code Blame 320 lines (237 loc) · 14.8 KB Raw 1 2 3 4 5 6 7 8 9 10 11 12 13 ...

22 lines (14 loc) · 556 Bytes main StableDAW / stable_audio_3 / verbose.py Top Code Blame 22 lines (14 loc) · 556 Bytes Raw 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 ...

303 lines (267 loc) · 14.3 KB main StableDAW / stable_audio_3 / interface / aeiou.py Top Code Blame 303 lines (267 loc) · 14.3 KB Raw 1 2 3 4 5 6 7 8 9 10 11 12 13 ...

Your prompt only affects the regeneration step. Default: OFF. Gives cleaner separation between "what to keep from the original" and "what to change with the prompt." Best for prompt-based restyling — …

Windows Troubleshooting Common issues and fixes specific to running Stable Audio 3 on Windows.