heartmula

Generate full songs from lyrics and tags using HeartMuLa models.

Updated Apr 2, 2026
One-click install
npx skills add https://github.com/JKhyro/HERMES-AGENT --skill heartmula-jkhyro
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: heartmula
Source: https://github.com/JKhyro/HERMES-AGENT/tree/main/skills/media/heartmula
Command: npx skills add https://github.com/JKhyro/HERMES-AGENT --skill heartmula-jkhyro

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

HeartMuLa removes the complexity of setting up open-source music generation so you can turn lyrics and style tags into full songs without assembling a custom audio pipeline.

Core Features & Use Cases

  • Song generation from text: Create complete music tracks from lyrics plus comma-separated tags.
  • Open-source Suno-like workflow: Use HeartMuLa, HeartCodec, HeartTranscriptor, and HeartCLAP for generation, reconstruction, transcription, and alignment.
  • Practical deployment guidance: Follow hardware, installation, checkpoint download, and compatibility notes to run locally on GPU or, more slowly, on CPU.

Quick Start

Ask the assistant to set up HeartMuLa, download the required checkpoints, and generate an MP3 from 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 multilingual songs from lyrics locally?

To generate multilingual songs from lyrics locally, you provide your lyrics and comma-separated style tags to the HeartMuLa model family. The Skill handles the installation, checkpoint downloads, and device configuration to output a complete music track.

What do I need to set up GPU-accelerated AI music generation?

Setting up GPU-accelerated AI music generation requires configuring your device settings and applying compatibility patches for transformers and codec loading. The Skill guides you through dependency upgrades and model checkpoint downloads to enable local GPU processing.

Does the HeartMuLa model family support CPU-only audio codec reconstruction?

Yes, the HeartMuLa model family supports CPU-only audio codec reconstruction through HeartCodec, but it runs more slowly than GPU-accelerated setups. The Skill provides practical deployment guidance for both hardware configurations.

Can I transcribe and align lyrics automatically during song generation?

Yes, you can transcribe and align lyrics automatically during song generation by utilizing HeartTranscriptor and HeartCLAP. These components handle lyric transcription and alignment alongside the primary music generation workflow.

What is the best way to turn text lyrics into a full MP3 track without assembling a custom pipeline?

The best way to turn text lyrics into a full MP3 track without assembling a custom pipeline is using an open-source Suno-like workflow. This Skill removes setup complexity by managing dependencies, checkpoints, and compatibility patches automatically.

Why do I need to apply compatibility patches for transformers and codec loading?

You need to apply compatibility patches for transformers and codec loading to ensure the open-source audio pipeline functions correctly with current library versions. The Skill automates these patches to prevent loading errors during model initialization.