heartmula

Name: heartmula
Availability: InStock
Author: cabezno

Generate MP3 songs from lyrics files and comma-separated tags using HeartMuLa with lazy loading enabled.

Updated May 15, 2026

Search Tags:#music generation #audio generation #multilingual #mp3 #lyrics #suno alternative #open-source model

Ready to run

npx skills add https://github.com/cabezno/bmb-encover-agent --skill heartmula-cabezno

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: heartmula
Source: https://github.com/cabezno/bmb-encover-agent/tree/main/skills/media/heartmula
Command: npx skills add https://github.com/cabezno/bmb-encover-agent --skill heartmula-cabezno

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes assets (resource) components.

What problem does it solve?

HeartMuLa solves the problem of turning written lyrics and stylistic tags into full, listenable songs without needing music production expertise.

Core Features & Use Cases

Lyrics + Tags Conditioned Generation: Produces complete songs from structured lyrics (with section markers) plus comma-separated tags.
Open-Source, Local Music Foundation Models: Uses HeartMuLa generation models together with HeartCodec for audio reconstruction, with optional transcription/alignment components in the model family.
Practical Hardware-Aware Workflow: Supports GPU acceleration and offers lazy-loading to reduce VRAM requirements while generating MP3 outputs.

Quick Start

Use this command: "Generate a full song MP3 from my lyrics file and my tags file using HeartMuLa with lazy loading enabled, saving the result to output.mp3."

Frequently Asked Questions about heartmula

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

FAQPage Schema

How do I generate songs from lyrics locally without using Suno?▼

You can generate full songs locally by converting structured lyrics with section markers and comma-separated style tags into MP3 audio using open-source models, avoiding external cloud services like Suno.

Does open-source music generation support multilingual lyrics and MP3 output?▼

Yes, open-source music generation supports multilingual lyrics inputs and outputs complete MP3 files by conditioning the generation model on your text and reconstructing the audio with a dedicated codec.

Can I run end-to-end audio generation on a GPU with limited VRAM?▼

Yes, you can run end-to-end audio generation on limited VRAM by enabling lazy loading settings in the generation script, which reduces memory requirements while still producing MP3 audio.

What is the best way to structure lyrics for AI music generation?▼

The best way to structure lyrics for AI music generation is to use section markers for different song parts and pair them with comma-separated style tags, allowing the model to accurately condition the generated audio.

Do I need to download specific checkpoints to generate audio locally?▼

Yes, you must download required model checkpoints and install the generation stack locally before running the provided script with appropriate GPU or CPU device flags to produce MP3 files.