Configure HeartMuLa music generation models for local MP3 song creation.
npx skills add https://github.com/graniet/kheish --skill heartmula-graniet
Please help me install this Agent Skill. Skill: heartmula Source: https://github.com/graniet/kheish/tree/main/skills/media/heartmula Command: npx skills add https://github.com/graniet/kheish --skill heartmula-graniet
This Skill eliminates the complex, error-prone manual work of setting up the open-source HeartMuLa music generation model family, which requires specific dependency fixes, source code patches, and hardware configuration to run correctly, making local AI music generation accessible without deep technical expertise.
Use the heartmula skill to generate a 2-minute upbeat rock song with the lyrics "[Verse] Driving down the highway, wind in my hair / [Chorus] This is the life, we don't have a care" and the tags "rock,energetic,guitar,drums,male-vocal", then save the output as a high-quality MP3 file.
High-intent search queries and answers about installing and using this skill.
To set up open-source AI music generation locally, you must clone the repository, configure a Python 3.10 virtual environment, resolve dependency conflicts, and apply source code patches for hardware compatibility. This process ensures the model functions correctly without standard installation errors.
Yes, you can generate AI music with low VRAM or a CPU-only system by utilizing lazy loading options and specific hardware configurations. The setup supports running the 3B model on systems with as little as 8GB of VRAM, ensuring broad hardware accessibility.
You can convert text lyrics and genre tags into full stereo audio by inputting custom lyrics and style tags into the generation workflow. This process produces high-fidelity 48kHz stereo MP3 songs, configurable by length, sampling, and audio quality parameters.
Yes, an open source alternative to Suno exists for generating multilingual AI songs. It enables local and offline creation of full songs from lyrics and genre tags, providing content creators and musicians with royalty-free audio outputs without commercial licensing restrictions.
The HeartMuLa music generation model requires source code patches to resolve non-standard dependency fixes and ensure compatibility with newer package versions. These modifications are necessary to bypass manual setup errors and execute the model correctly across different hardware environments.
m13vConfigure ai-browser-profile to extract browser data from local files on macOS.
calesthioGenerate music, vocal covers, and isolated stems via RunPod ACE-Step.
spacecat2002Select SGLang Diffusion configuration flags for generation speed and GPU memory usage.
mietheGenerate song lyrics with rhyme, meter, syllable, and profanity constraints from pinned MCP sources.
