https://whop.com/audiolab-0d48/free-starter-stable-audio-3-background-music-studio
> FULL PACK ($45): https://whop.com/audiolab-0d48/stable-audio-3-background-music-studio-beds-loops — use code STAB330 for 30% off ($45 → $31.50).
## FREE vs FULL
| | FREE (this download) | FULL ($45) |
|---|---|---|
| Workflows | 1 (text-to-bed loop, 90 s) | 4 |
| Prompt library | 5 starter prompts | 30 prompts (genre / mood / energy) |
| Long-form bed (240 s) | — | Yes (Medium) |
| Inpaint / extend | — | Yes |
| Loop-seaming recipe | — | Yes (bar math + ffmpeg) |
| Ducking-under-voiceover guide | — | Yes |
| Guides (EN + AR) | Quick README only | Yes |
## What this is
Stable Audio 3 is Stability AI's open-weights music family. A November 2026 guide named the Medium model "Best overall for background music", and Stability's own 12 Nov 2026 explainer lists Stable Audio 3.0 as the only open-weights family on its list — because it makes instrumental music and sound effects and generates no intelligible vocals or lyrics. That is exactly what you want under a voiceover. This free lite kit gives you one working text-to-background-music workflow (Medium, 90 s, stereo 44.1 kHz) and 5 starter prompts so you can generate your first bed today. The full pack adds the Small-Music CPU loop, the 240 s Medium long-form bed, inpainting/continuation, loop-seaming, the full 30-prompt genre library and the ducking guide. Honest caveat: the model itself released in May 2026 — this is a November coverage/demand pick, not a new release.
## Files you need (not included)
stable_audio_3_medium.safetensors in models/checkpoints/ and t5gemma_b_b_ul2.safetensors in models/text_encoders/ (Comfy-Org/stable-audio-3 on Hugging Face — accept the licence agreement first). Update ComfyUI to the latest release and open the Stable Audio 3.0 Medium template. No model files, no keys included.
Licence: Stability AI Community License — free commercial use under $1M annual revenue; above $1M a separate licence is required. Keep the NOTICE file and display "Powered by Stability AI". Keep the prompt, seed and model version for every track you publish. This is not legal advice.
Built from public research; all content original. No brand assets, no credentials. Figures cited are from the sources in ADAPTED_NOTES.md — they describe the market, not a promise. Structure-verified only — not GPU-tested.
SFW only. 'Depicts an actual person' = No. No 18+ attestation.
Full pack on Whop: https://whop.com/audiolab-0d48/stable-audio-3-background-music-studio-beds-loops

