La' Musica

Local music generation app powered by ACE-Step 1.5.

Generate music from text descriptions, create song lyrics in multiple languages, and manage your generation history — all running locally on your GPU.

Architecture

Service	Tech	Port	Description
frontend	React + Vite + Tailwind	3000	Web UI for music generation
backend	Python FastAPI	8000	API middleware, business logic
acestep	ACE-Step 1.5 API	8001	Music generation (GPU)
ollama	Ollama + Qwen2.5-7B	11434	Song lyrics generation (CPU)
postgres	PostgreSQL 16	5432	Generation history storage

Models (optimized for 16 GB VRAM)

• DiT: acestep-v15-turbo — 8 inference steps, very high quality
• LM: acestep-5Hz-lm-1.7B — vLLM backend, great for 12-16 GB GPU
• Lyrics: qwen2.5:7b — multilingual (EN/RU/BY), runs on CPU via Ollama

Quick Start

Prerequisites

• Docker + Docker Compose v2
• NVIDIA GPU with 16 GB VRAM
• NVIDIA Container Toolkit

Launch

# 1. Clone the repo
git clone https://github.com/W1neSkin/ACE-Step-Music.git
cd ACE-Step-Music

# 2. Copy and edit env config
cp .env.example .env
# (Optional) edit .env to customize DB password, model settings, etc.

# 3. Start all services
docker compose up --build

# 4. Open the app
# Frontend: http://localhost:3000
# Backend API: http://localhost:8000/docs

First launch will take 10-20 minutes to download models (cached for subsequent runs).

Development (without Docker)

Backend:

cd backend
pip install -r requirements.txt
uvicorn app.main:app --reload --port 8000

Frontend:

cd frontend
npm install
npm run dev

Features

• Music Generation — describe a style, write lyrics, set BPM/key/duration, generate multiple variants
• Lyrics AI — generate song lyrics in English, Russian, or Belarusian using Qwen 2.5
• Audio Player — waveform visualization, variant switching, download
• History — browse and replay past generations
• Dark theme — modern UI inspired by Suno/Spotify aesthetic

API Endpoints

Method	Path	Description
POST	/api/music/generate	Submit generation task
GET	/api/music/status/{task_id}	Poll task status
GET	/api/music/audio?path=...	Download generated audio
GET	/api/music/models	List available models
POST	/api/lyrics/generate	Generate song lyrics
GET	/api/history	List generation history
GET	/api/history/{id}	Get generation details
DELETE	/api/history/{id}	Delete a generation
GET	/api/health	Health check

License

MIT