heartmula

Name: heartmula
Availability: InStock
Author: hungthinh04

Generate full songs from lyrics and style tags with HeartMuLa.

Updated May 12, 2026

Search Tags:#music generation #audio synthesis #lyrics #heartmula #offline inference #song creation

Ready to run

npx skills add https://github.com/hungthinh04/Hermes_AI_Agent --skill heartmula-hungthinh04

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: heartmula
Source: https://github.com/hungthinh04/Hermes_AI_Agent/tree/main/skills/media/heartmula
Command: npx skills add https://github.com/hungthinh04/Hermes_AI_Agent --skill heartmula-hungthinh04

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

HeartMuLa removes the need to manually produce music by turning lyrical ideas and descriptive tags into complete, listenable songs with an open-source workflow.

Core Features & Use Cases

Lyrics-to-song generation: Create full tracks from structured lyrics plus comma-separated style tags.
Open-source music stack: Work with HeartMuLa, HeartCodec, HeartTranscriptor, and HeartCLAP for generation, audio reconstruction, and transcription tasks.
Practical deployment: Use it for local/offline song creation, GPU-accelerated inference, or troubleshooting model setup and compatibility issues.

Quick Start

Ask the skill to generate a song from your lyrics and style tags, then save the finished track as an MP3.

Frequently Asked Questions about heartmula

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

FAQPage Schema

How do I generate full songs from lyrics and style tags?▼

To generate full songs from lyrics, provide structured text and comma-separated style tags. The system processes your inputs locally to create listenable tracks using lyrics-conditioned generation.

Can I run music generation locally offline without an internet connection?▼

Yes, you can run music generation offline. This workflow supports local inference on either CPU or GPU-accelerated CUDA devices, allowing you to create music completely offline.

Do I need model checkpoints for offline audio synthesis?▼

Yes, model checkpoints are required for offline audio synthesis. You must configure these checkpoints and select your CUDA or CPU device to ensure the generation and audio reconstruction tasks execute properly.

How do I fix dependency conflicts and transformer cache loading issues?▼

To fix dependency conflicts and transformer cache loading issues, apply the provided compatibility fixes. These fixes resolve setup errors during local model initialization and ensure smooth offline inference.

What is the best way to transcribe and reconstruct audio for song creation?▼

The best way to transcribe and reconstruct audio is using the open-source HeartTranscriptor and HeartCodec stacks. They handle transcription tasks and audio reconstruction to support your full song creation workflow.