HeartMuLa/heartlib

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

3749 stars · 450 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

HeartMuLa is a family of open-source music foundation models that generate music conditioned on lyrics and tags with multilingual support. The repo also includes HeartCodec (a music codec), HeartTranscriptor (lyrics transcription), and HeartCLAP (audio-text alignment for cross-modal retrieval).

Use cases

• generate songs from lyrics and style tags
• create multilingual AI music with controllable lyrics
• transcribe lyrics from audio recordings
• encode and decode music with a high-fidelity neural codec
• build cross-modal music search with audio-text embeddings
• benchmark AI music generation models

When to choose

• you need open-source, lyrics-conditioned music generation with multilingual support
• you want state-of-the-art open music quality comparable to commercial services
• you need companion tools for lyrics transcription or music retrieval embeddings

When to avoid

• you need lightweight real-time music synthesis on CPU
• you only need simple MIDI or symbolic music generation
• you lack GPU resources for large 3B+ parameter models

Facets

library · maturity active

machine-learning deep-learning audio-processing llm-inference speech-recognition artificial-intelligence deep-learning python cross-platform music-generation music-codec lyrics-to-song text-to-music audio-text-alignment lyrics-transcription foundation-model audio natural-language-processing gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="HeartMuLa/heartlib")