heartmula

Generate music from multilingual lyrics and tags on local hardware.

Updated May 2, 2026
One-click install
npx skills add https://github.com/AlvaroBiano/hermes-agent --skill heartmula-alvarobiano
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: heartmula
Source: https://github.com/AlvaroBiano/hermes-agent/tree/main/skills/media/heartmula
Command: npx skills add https://github.com/AlvaroBiano/hermes-agent --skill heartmula-alvarobiano

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

HeartMuLa enables local, open-source music generation conditioned on user-provided lyrics and tags, delivering end-to-end audio without relying on hosted services.

Core Features & Use Cases

  • Open-source foundation models: HeartMuLaGen, HeartCodec, HeartTranscriptor, and HeartCLAP for end-to-end music creation.
  • Lyrics + tags conditioned generation: Create songs by combining multilingual lyrics with descriptive tags.
  • Use Case: A songwriter generates a demo track from a set of lyrics and mood/genre tags for rapid prototyping.
  • Offline/local execution: Suitable for on-device generation with configurable model components.

Quick Start

Start a basic generation task by providing lyrics and a set of tags to produce an audio track.

Frequently Asked Questions about heartmula

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

FAQPage Schema
How do I generate music from lyrics and tags locally?▼

To generate music from lyrics and tags locally, you provide multilingual text and descriptive genre or mood tags to the model, which then produces an end-to-end audio track entirely on your device without relying on hosted services.

Can I generate music offline using CPU or GPU environments?▼

Yes, you can generate music offline using configurable model components across CPU or GPU environments. The modular stack allows you to initialize models and codecs on local hardware, ensuring on-device generation without internet dependency.

What is lyrics and tags conditioned music generation?▼

Lyrics and tags conditioned music generation is a process where foundation models create audio tracks by interpreting multilingual text alongside descriptive mood or genre tags. This enables creators to rapidly prototype demo tracks from textual descriptions.

Do I need specialized codecs to initialize open-source music models?▼

Yes, you need to download and install specific codecs alongside foundation models to initialize the open-source music generation stack. These codecs are essential modular components required to process and output the final audio.

What are the limitations of local music generation compared to hosted services?▼

Local music generation limitations depend on your hardware capabilities, requiring sufficient CPU or GPU resources to run the modular stack. Unlike hosted services, all processing occurs on-device, meaning generation speed is bound by your local compute power.