https://whop.com/audiolab-0d48/free-starter-moss-soundeffect-v2-game-sfx-studio
> FULL PACK ($45): https://whop.com/audiolab-0d48/moss-soundeffect-v2-game-sfx-studio-text-to-sfx-library — use code MOSS30 for 30% off ($45 → $31.50).
## FREE vs FULL
| | FREE (this download) | FULL ($45) |
|---|---|---|
| Workflows | 1 (short-clip SFX, 3 s) | 4 (short-clip + 30 s long-form + batch + ambience bed) |
| Prompt library | 6 prompts | 48 prompts (UI, combat, footsteps by surface, nature, animals, machines, magic/sci-fi, ambience) |
| Batch generation | — | Yes (4 sounds, one graph) |
| Ambience bed (30 s + accent) | — | Yes |
| Loudness / LUFS finishing notes | — | Yes (−16 / −24 / −28 LUFS game targets) |
| Guides (EN + AR) | Lite README only | Yes (quickstart, Arabic guide, doctor, test report) |
| Cheat sheet | — | Yes (Whop free starter only) |
## What this is
MOSS-SoundEffect v2.0 (OpenMOSS, released 26 May 2026) is a 1.3B text-to-SFX model: type a sentence, get a 48 kHz mono sound effect back, up to 30 seconds, from English or Chinese prompts. Footsteps, impacts, UI clicks, ambience, nature, animals — effects only, no speech, no music. Its licence is Apache 2.0, fully commercial-safe, which is why a November 2026 guide named it the "Best commercial-friendly text-to-SFX" model — and a 16 Nov 2026 game-dev tutorial uses it to make every sound effect for a Roblox game. This free lite kit gives you one working short-clip workflow (footsteps on gravel, 3 s) with 6 starter prompts so you can hear your first sound today. The full pack adds the 30 s long-form workflow, batch generation, the ambience bed, all 48 game-audio prompts and the loudness finishing notes.
## How it runs (read before downloading)
MOSS does NOT run inside ComfyUI's Python. It runs in a separate Python 3.12 worker (own venv, torch 2.9) that the nodes talk to over JSON lines. Install the node pack comfyui-moss-soundeffect-v2, HuizhiXu) from ComfyUI Manager (search "MOSS SoundEffect v2"), download the weights yourself OpenMOSS-Team/MOSS-SoundEffect-v2.0, ~11 GB) into ComfyUI/models/moss-soundeffect-v2/MOSS-SoundEffect-v2.0 (must contain model_index.json), and let the Loader node build its own .venv on first run. The worker holds ~11 GB VRAM while resident — use the Unload node (wired in the workflow) to free it. First generation compiles the DiT and can take minutes with progress stuck at 0 — it is not hung.
## Files you need (not included)
No model files included. You download the weights yourself from Hugging Face (see above). No API keys — none are used, none are included.
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.
SFW only. 'Depicts an actual person' = No. No 18+ attestation.
Full pack on Whop: https://whop.com/audiolab-0d48/moss-soundeffect-v2-game-sfx-studio-text-to-sfx-library

