Generate music from lyrics and tags using HeartMuLa models on local hardware.
npx skills add https://github.com/ever-oli/io --skill heartmula-ever-oli
Please help me install this Agent Skill. Skill: heartmula Source: https://github.com/ever-oli/io/tree/main/skills/media/heartmula Command: npx skills add https://github.com/ever-oli/io --skill heartmula-ever-oli
HeartMuLa enables offline, open-source music generation anchored on lyrics and tags, removing reliance on proprietary services to produce songs locally.
Install the required environment, download the HeartMuLa checkpoints, and run the provided example script to generate music from lyrics and tags.
High-intent search queries and answers about installing and using this skill.
You can generate music from lyrics locally by running the HeartMuLa models on GPU-enabled hardware. This open-source approach removes reliance on proprietary services by using downloaded checkpoints and a Python environment for offline song creation.
Offline AI music generation requires CUDA-enabled PyTorch, downloaded model checkpoints, and a Python environment configured with the run script. A GPU-enabled machine is necessary to support the local processing and fast turnaround.
Yes, you can create songs using multilingual lyrics with open-source AI. The generation process supports multiple languages and provides fine-grained control over the musical output through descriptive tags.
Lyric-driven music generation works by processing text lyrics alongside descriptive tags using models like HeartMuLaGen and HeartCLAP. These tags provide fine-grained control, guiding the local AI to shape the song's style and structure.
The best way to set up a local environment for lyric-driven song creation is to install CUDA-enabled PyTorch, download the HeartMuLa checkpoints, and configure the provided Python run script with optional lazy_load and device settings.
A key limitation of generating music from lyrics on local hardware is the strict requirement for CUDA-enabled PyTorch and a compatible GPU. Without these, the open-source models cannot process the generation workload offline.
mietheGenerate song lyrics with rhyme, meter, syllable, and profanity constraints from pinned MCP sources.
KaiserWhoLearnsExtract, normalize, and mix audio from video files using FFmpeg commands.
juliuss1907Generate music from lyrics and tags using open-source models.
Simon-Copilot-StudioGenerate full songs from lyrics and tags using HeartMuLa models.
