---

name: mmaudio description: "MM Audio. Generate synchronized audio from video content using MM Audio. Triggers: mmaudio, video generation" allowed-tools: Bash(curl *), WebFetch

MM Audio

Generate synchronized audio from video content using MM Audio. Supports configurable duration, negative prompt support, reproducible results via seed.

Quick Start

Requires an each::labs API key. Get one at eachlabs.ai.

curl -X POST https://api.eachlabs.ai/v1/prediction \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "model": "mmaudio",
    "version": "0.0.1",
    "input": {
      "video": "https://example.com/video.mp4",
      "prompt": "Birds chirping in a forest with a gentle stream"
    }
  }'

Parameters

Parameter	Type	Default	Description
cfg_strength	number	4.5	Guidance strength (CFG)
duration	number	8	Duration of output in seconds
negative_prompt	string	music	Negative prompt to avoid certain sounds
num_steps	integer	25	Number of inference steps
prompt	string		Text prompt for generated audio
seed	integer		Seed
video	string		Optional video file for video-to-audio generation

Examples

Generate audio for video:

curl -X POST https://api.eachlabs.ai/v1/prediction \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "model": "mmaudio",
    "version": "0.0.1",
    "input": {
      "video": "https://example.com/video.mp4",
      "prompt": "Birds chirping in a forest with a gentle stream"
    }
  }'

Urban soundscape:

curl -X POST https://api.eachlabs.ai/v1/prediction \
  -H "Content-Type: application/json" \
  -H "X-API-Key: $EACHLABS_API_KEY" \
  -d '{
    "model": "mmaudio",
    "version": "0.0.1",
    "input": {
      "video": "https://example.com/city-video.mp4",
      "prompt": "Busy city traffic sounds with honking and people talking",
      "seed": 42,
      "duration": 10
    }
  }'

Documentation

• each::labs Docs
• API Reference