741,000+ skills indexed

luckybbjason1COMMUNITY

Explore all skills by this author →

Verified:GitHub

heartmula

Generates full songs from lyrics and style tags using HeartMuLa open-source music models.

Updated Sep 9, 2026

Search Tags:#music generation #open-source #text-to-music #heartmula #lyrics to song #ai audio #gpu inference

One-click install

npx skills add https://github.com/luckybbjason1/trading --skill heartmula-luckybbjason1

Or copy as Structured Prompt for Agent▼

Please help me install this Agent Skill.
Skill: heartmula
Source: https://github.com/luckybbjason1/trading/tree/main/.hermes/skills/media/heartmula
Command: npx skills add https://github.com/luckybbjason1/trading --skill heartmula-luckybbjason1

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Generating complete songs with vocals from text requires either paid services like Suno or complex local model setups. This Skill installs and runs HeartMuLa, an open-source music generation model family, so you can produce full songs locally from lyrics and style tags. ## Core Features & Use Cases - Song Generation from Lyrics + Tags: Produces MP3 audio (48kHz stereo) conditioned on bracketed lyrics structure and comma-separated style tags using 3B/7B models. - Local and Offline Operation: Runs on your own GPU with as little as 8GB VRAM using lazy loading, with multi-GPU splitting support. - Complete Pipeline: Includes HeartCodec for high-fidelity audio reconstruction, HeartTranscriptor for lyrics transcription, and HeartCLAP for audio-text alignment. - Use Case: A musician wants an open-source Suno alternative to prototype song ideas locally—write lyrics with [Verse]/[Chorus] tags, specify tags like "piano,happy,wedding", and generate a 4-minute track in about 4 minutes. ## Quick Start Ask the assistant to install HeartMuLa and generate a song from your lyrics file and style tags, saving the result 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 a song from lyrics with HeartMuLa?▼

Run the run_music_generation.py example with --lyrics and --tags files, a model path, and an output path. Lyrics use bracketed structure tags like [Verse] and [Chorus]; tags are comma-separated style descriptors like "piano,happy,wedding".

What GPU do I need for local AI music generation?▼

HeartMuLa needs a minimum of 8GB VRAM with lazy loading enabled, and 16GB or more is recommended. The 3B model peaks at about 6.2GB VRAM with lazy_load, and you can split models across two GPUs with --mula_device and --codec_device.

Can HeartMuLa run on CPU or without an NVIDIA GPU?▼

Yes, CPU mode works via --mula_device cpu --codec_device cpu, but generation is extremely slow (30-60+ minutes per song) and needs 12GB+ RAM. Without a GPU, a cloud GPU service or the online demo at heartmula.github.io is recommended.

Why does HeartMuLa fail after installing dependencies?▼

The pinned dependencies conflict with newer packages, so you must upgrade datasets and transformers after install. Two source patches are also required: RoPE cache reinitialization for transformers 5.x and ignore_mismatched_sizes=True for HeartCodec loading.

What audio quality and length does HeartMuLa output?▼

Output is MP3 at 48kHz stereo, 128kbps, with a default maximum length of 240 seconds (4 minutes). Generation runs at roughly real-time, so a 4-minute song takes about 4 minutes on GPU.

What are the limitations of HeartMuLa music generation?▼

Style tags may be ignored since lyrics tend to dominate conditioning, and bf16 degrades HeartCodec audio quality so fp32 is required. Triton is unavailable on macOS, and RTX 5080 incompatibility has been reported upstream.

Related Skills

Explore More Design & Creative Skills→

suno

Write Suno AI music prompts with style tags, metatag song structures, and lyrics.

video-production-buddy

acestep

Generate custom audio assets and separate stems for video production.

cosmicstack-labs

audio-extraction

Extract and convert audio from video sources into MP3, FLAC, AAC, Opus, or WAV.

heartmula

Generate music and songs from lyrics and tags using HeartMuLa models.