Author: ace_step Version: 0.0.1 Type: tool
A Dify plugin for AI music generation powered by ACE-Step v1.5. Generate high-quality music from text descriptions and optional lyrics.
- Text to Music: Generate music by describing the desired style, mood, instruments, and genre
- Lyrics Support: Provide structured lyrics with tags like
[Verse],[Chorus],[Bridge] - Multi-language Vocals: Supports English, Chinese, Japanese, and Korean vocals
- Cover & Repaint: Additional task types for song covers and audio repainting
- Fine-grained Control: Configure duration, BPM, temperature, guidance scale, and more
- Install this plugin in your Dify instance
- Configure the Server URL (default:
https://api.acemusic.ai) - Optionally provide an API Key if your server requires authentication
| Parameter | Type | Default | Description |
|---|---|---|---|
| Music Description | string | (required) | Describe the music style you want |
| Lyrics | string | (optional) | Song lyrics with structure tags |
| Task Type | select | text2music | text2music / cover / repaint |
| Vocal Language | select | en | en / zh / ja / ko |
| Duration | number | 30 | Audio duration in seconds (5-300) |
| Instrumental | boolean | false | Generate without vocals |
| BPM | number | - | Beats per minute (60-300) |
| Seed | string | -1 | Random seed (-1 = random) |
| Thinking | boolean | true | Enable LM thinking mode |
| Temperature | number | 0.85 | Sampling temperature (0-2) |
| Guidance Scale | number | 7.0 | DiT guidance scale (0-20) |
| Inference Steps | number | 8 | Diffusion inference steps (1-200) |
- Text: Generated music metadata and description
- JSON: Structured metadata (model, content, has_audio)
- Audio: MP3 audio file as blob