Windows · Local AI · ComfyUI

Make music locally.
Keep everything yours.

Fenix Music Studio is a standalone Windows music creation app built around local AI generation. Create with ACE-Step 1.5 or MiniMax Music 3, organise your tracks, edit, master and manage projects without sending your library to a cloud service.

Download for Windows · v0.9.5 Setup ComfyUI & models

Windows 10/11 · NVIDIA recommended · ComfyUI required for generation

ACE-Step 1.5 MiniMax Music 3 ComfyUI Local Library Built-in DAW Mastering

Built for music

One app for creating, managing and finishing tracks.

Fenix Music Studio keeps the whole music workflow in one focused desktop app, from the first prompt to the finished audio file.

♪

Local AI generation

Generate complete tracks through your own ComfyUI install using ACE-Step 1.5 or MiniMax Music 3.

Simple + advanced creation

Start quickly with streamlined controls, then open deeper generation options when you want full control.

▦

Music library

Keep generated and imported tracks indexed locally with search, favourites, projects and metadata.

≋

DAW workflow

Move from generation into editing without leaving the app, with project-based audio tools built around the music.

Styles & ACE LoRAs

Organise reusable styles and use compatible ACE-Step LoRAs stored in your normal ComfyUI model folders.

Mastering & conversion

Finish tracks inside the studio. Optional FFmpeg support expands mastering, conversion and audio analysis features.

Generation queue

Queue work instead of babysitting each generation and keep track of what is waiting, running or complete.

Projects

Group tracks, versions and related work together instead of dumping everything into one endless output folder.

⌁

Your files stay normal

Audio remains on disk and the app indexes it locally. Your tracks are not locked inside a proprietary cloud library.

Workflow

From idea to finished track.

No complicated platform stack. Fenix Music Studio stays focused on the actual music workflow.

CreateWrite your prompt, lyrics, style and generation settings.

GenerateSend the job to ACE-Step 1.5 or MiniMax Music 3 through ComfyUI.

EditMove into the built-in DAW and shape the track.

OrganiseStore versions, favourites and projects in the local library.

MasterPrepare the finished result and export it where you want it.

AI engines

ACE-Step 1.5 and MiniMax Music 3.

Both generation workflows are supplied internally by Fenix Music Studio. Normal users do not need to download, import or edit workflow JSON files.

Included workflow

ACE-Step 1.5

A flexible local music generation engine with native ComfyUI support and optional LoRA loading.

Turbo diffusion model
0.6B and 4B ACE text encoders
ACE 1.5 audio VAE
Optional ACE LoRAs from models/loras/ace/

Included workflow

MiniMax Music 3

Higher-end local music generation with a recommended FP16 model and an INT8 alternative for lower-VRAM systems.

FP16 diffusion model recommended
INT8 diffusion alternative available
Pruned INT8 text encoder
MiniMax audio VAE / DAV

System requirements

Built for a local Windows AI setup.

Fenix Music Studio itself is lightweight. Model execution is handled by your existing ComfyUI installation.

Minimum

Operating systemWindows 10 / 11, 64-bit

GPUNVIDIA CUDA-capable

VRAM12 GB

System RAM32 GB

Storage~20 GB free per model engine

RuntimeMicrosoft Edge WebView2

Install ComfyUI. Add the models. Start making music.

Fenix Music Studio connects to ComfyUI over its local HTTP API. CUDA, PyTorch, Python, GPU drivers and model execution remain entirely inside your ComfyUI installation.

1. Install or update ComfyUI

Use an up-to-date ComfyUI build. No third-party custom-node package is required. Fenix uses native comfy-core audio nodes for ACE-Step 1.5 and MiniMax Music 3.

ComfyUI should normally be available at:

http://127.0.0.1:8188

A different address or port can be configured later inside Fenix Music Studio Settings.

2. Install ACE-Step 1.5 models

Place these four files into the matching folders inside your ComfyUI installation:

ComfyUI/
└── models/
    ├── diffusion_models/
    │   └── acestep_v1.5_turbo.safetensors
    ├── text_encoders/
    │   ├── qwen_0.6b_ace15.safetensors
    │   └── qwen_4b_ace15.safetensors
    └── vae/
        └── ace_1.5_vae.safetensors

ACE-Step 1.5 Turbo diffusion modelDownload ↗ Qwen 0.6B ACE text encoderDownload ↗ Qwen 4B ACE text encoderDownload ↗ ACE-Step 1.5 VAEDownload ↗

Optional ACE LoRAs can be stored in:

ComfyUI/models/loras/ace/

Subfolders are supported. The recommended ACE model set is approximately 14.7 GB.

3. Install MiniMax Music 3 models

The recommended MiniMax Music 3 setup uses the FP16 diffusion model:

ComfyUI/
└── models/
    ├── diffusion_models/
    │   └── minimax_music3_dit_fp16.safetensors
    ├── text_encoders/
    │   └── minimax_music3_text_encoder_pruned_int8_convrot.safetensors
    └── vae/
        └── minimax_music3_dav.safetensors

MiniMax Music 3 FP16 diffusion modelDownload ↗ MiniMax Music 3 pruned INT8 text encoderDownload ↗ MiniMax Music 3 audio VAE / DAVDownload ↗ Alternative: MiniMax Music 3 INT8 diffusion modelLower VRAM ↗

The recommended MiniMax model set is approximately 14.3 GB. Fenix automatically detects compatible MiniMax filenames.

4. Start ComfyUI, then launch Fenix Music Studio

Start ComfyUI before generating music. Fenix Music Studio will connect to the local API and use its built-in ACE-Step 1.5 and MiniMax Music 3 workflows automatically.

You do not need to import workflow JSON files for normal use.

5. Optional: install FFmpeg for extra audio features

FFmpeg is not required to generate or play tracks. It is optional and used for extra mastering, conversion and extended audio-analysis functions.

Fenix Music Studio 0.9.5

Your local AI music workspace.

Generate with your own hardware, keep the files on your own machine, and manage the full music workflow from one Windows app.

Download FenixMusicStudio.exe Read setup instructions