# MMAudio - Core.Today AI API

> Add AI-generated sound to any video (5M+ runs) — synthesizes synchronized audio (ambience, effects, foley) from the video content and an optional text prompt. The perfect finisher for silent AI-generated clips.

- **Provider**: MMAudio
- **Model ID**: zsxkib/mmaudio
- **Category**: Video Generation
- **Credits**: 11 per generation
- **Speed**: Medium
- **Quality**: High

## Features
- Video-synchronized audio synthesis
- Optional prompt to steer the sound
- Negative prompt (default excludes music)
- Also works from a still image + duration

## Use Cases
- Adding sound to AI-generated videos
- Foley for short clips
- Ambience for product videos
- Silent archive footage revival

## API Endpoint

Base URL: https://api.core.today/v1
Create Prediction: POST /predictions
Get Status: GET /predictions/{job_id}

## Authentication
Header: X-API-Key: YOUR_API_KEY

## Input Parameters

### Optional
- **prompt**: string - Describe the sound you want (e.g. 'rain on leaves, distant thunder')
- **negative_prompt**: string (default: music) - Sounds to avoid (default avoids music)
- **video**: string - Video to add sound to
- **duration**: number (default: 8) - Audio duration in seconds
- **num_steps**: integer (default: 25) - Number of inference steps
- **cfg_strength**: number (default: 4.5) - Guidance strength (CFG)
- **seed**: integer - Random seed (-1 for random)
- **image**: string - Optional image file for image-to-audio generation (experimental)

## Examples

### Sound for an AI Clip
Add synchronized ambience to a generated video
```json
{
  "model": "zsxkib/mmaudio",
  "input": {
    "video": "https://example.com/generated-clip.mp4",
    "prompt": "ocean waves, seagulls, gentle wind"
  }
}
```

## Response Format
```json
{
  "job_id": "abc123",
  "status": "pending | processing | completed | failed",
  "result": "URL or data (when completed)"
}
```

## Usage Flow
1. POST /predictions with model and input -> receive job_id
2. GET /predictions/{job_id} -> poll until status is completed or failed
3. Result contains output URL(s)

## Tips
- Chain after Kling/Veo/Seedance generations — silent clips become alive for 15 credits
- Leave negative_prompt as 'music' for realistic foley; clear it if you want musical texture
- Describe 2-3 concrete sound sources in the prompt for the best sync

## Documentation
https://replicate.com/zsxkib/mmaudio