MMAudio_safetensors

MMAudio_safetensors

Provides MMAudio models in the safetensors format for Kijai's ComfyUI-MMAudio nodes, which add generated sound to video. For ComfyUI users who work with audio and video.

Made by Kijai on Hugging Face. These notes are Mindket's, written from its model page and model card; the model's own files and license are on Hugging Face.

About

What it is: This repository holds MMAudio models saved in the safetensors format by Kijai. The card names the MMAudio project on GitHub as the source of the models. It says they are meant to be used with Kijai's ComfyUI-MMAudio nodes. The MMAudio project is about making sound that fits a video, and the card itself gives no more detail.

Who it suits: It suits ComfyUI users who want to try MMAudio inside a node workflow, and researchers who want the models in a ready-to-load format.

Industry | Who uses it | How they use it | Example
Media | Video creators | Add generated sound to a clip inside ComfyUI | A silent clip that gets a sound track
Software | ComfyUI users | Load the models in the ComfyUI-MMAudio nodes | A workflow with an MMAudio node in it
Research | ML researchers | Try the MMAudio project's models in a ready node pack | Compare results from different settings
Content | Hobbyists | Experiment with sound for generated video | Add audio to a short test clip

If you do not use ComfyUI, you can skip it, since the card says these files are for the ComfyUI-MMAudio nodes.

Where to use it:

• ComfyUI, with the ComfyUI-MMAudio custom nodes
• Windows, macOS or Linux, where you download the files with Python

Use case | Who | What they get back
Sound for video | Video creators | Audio made by an MMAudio node in your workflow
Model testing | Researchers | MMAudio models in the safetensors format
Workflow building | ComfyUI users | A model set the node pack can load

How to set it up: You need Python with pip, ComfyUI, and Kijai's ComfyUI-MMAudio node pack. The card lists no file names or sizes, so check the file list on the page and the node pack's own notes.

• Python and pip
• ComfyUI
• The ComfyUI-MMAudio node pack, linked from the card
• Disk space for the files you choose

1. Install the Hugging Face client with pip.
2. Download the repository with the hf command.
3. Put the files where the node pack expects them, as its own page describes.

pip install -U huggingface_hub
hf download Kijai/MMAudio_safetensors

How to use it: Add the MMAudio nodes to a ComfyUI workflow and select the downloaded files in them. The card gives no run steps, so the node pack's page is the guide.

Example: You install the node pack in ComfyUI, download these files and pick them in the MMAudio nodes. The card gives nothing more, so the node pack's page covers the rest.

For newcomers and experts: Think of it as the model files for a plug-in. You do not run them alone, since the ComfyUI node pack loads them. The first small thing to try is the download command above.

This card has no details on the design, sizes or settings of the models. The source project's page is the place to read those.

MMAudio project, the source of the models
Models saved in safetensors format by Kijai
Loaded by the ComfyUI-MMAudio nodes
Sound generated inside your ComfyUI workflow

Good to know:

• Short card: it names the source project and the node pack, and gives no description of the files.
• Source: the models come from the MMAudio project on GitHub, so read its page for what they do.
• Popularity: 91 likes, and no downloads were recorded in the last 30 days.
• License: MIT, as the License line says. Check the source project's own terms too.
• Cost to run: no API key, account or paid service is named.
• Data it reaches: the videos and prompts you use in ComfyUI, and Hugging Face for the download.
• Not for: the card names no forbidden uses. Use only footage you have the right to use.

On Hugging Face, 13 November 2026

• Downloads in the last 30 days: 0
• Likes: 91
• License: mit

Links

• Model: https://huggingface.co/Kijai/MMAudio_safetensors
• Files: https://huggingface.co/Kijai/MMAudio_safetensors/tree/main

Get the model

Download its files with Hugging Face's command line:

pip install -U huggingface_hub
hf download Kijai/MMAudio_safetensors

Then follow the usage in its model card. Nothing in this kit runs by itself.