heartmula

Generate music locally from lyrics and tags using HeartMuLa models.

2.8k|332|Updated Jan 29, 2026
One-click install
npx skills add https://github.com/moltis-org/moltis --skill heartmula-moltis-org
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: heartmula
Source: https://github.com/moltis-org/moltis/tree/main/crates/skills/src/assets/media/heartmula
Command: npx skills add https://github.com/moltis-org/moltis --skill heartmula-moltis-org

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes assets (resource) components.

What problem does it solve?

HeartMuLa is a family of open-source music foundation models that generates music conditioned on lyrics and tags, designed to run locally without cloud services. It includes multilingual support to transform text into audio on your own hardware.

Core Features & Use Cases

  • HeartMuLaGen models generate music from lyrics and tags; HeartCodec provides high-fidelity audio reconstruction; HeartTranscriptor handles lyrics transcription; HeartCLAP enables audio-text alignment.
  • Local, offline generation on CPU or GPU hardware with optional lazy loading to save VRAM.
  • Use cases include music prototyping, educational demonstrations, and private music production without external dependencies.

Quick Start

Create a Python virtual environment, install dependencies, and run the provided music generation script with your lyrics and tags.

Frequently Asked Questions about heartmula

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate music locally from lyrics using open-source models?

Generate music locally from lyrics by running HeartMuLa models in a Python environment. You provide text lyrics and tags, and the models render the audio directly on your CPU or GPU hardware without requiring cloud services.

Can I run local audio generation on hardware with limited VRAM?

Local audio generation supports hardware with limited VRAM by utilizing optional lazy-loading. This feature manages memory efficiently, allowing you to run the music generation models on your own GPU or CPU setup.

What do I need to set up offline music generation from text?

Offline music generation from text requires a Python virtual environment, CUDA-enabled PyTorch, HuggingFace CLI tools, and HeartMuLa model checkpoints. Once installed, you can render audio completely offline.

Does local music generation support multilingual lyric input?

Local music generation does support multilingual lyric input. You can provide lyrics in various languages alongside tags, and the system will transform the text into audio on your device.

What is the best way to create private music prototypes without external dependencies?

The best way to create private music prototypes without external dependencies is using local generation models. This approach processes lyrics and tags entirely on your hardware, ensuring privacy for your music production or educational demonstrations.