ASLP-lab/DiffRhythm

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

2339 stars · 275 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

DiffRhythm is an open-source latent diffusion model for end-to-end full-length song generation, supporting lyrics-to-song, text style prompts, and instrumental music generation. It is the first open-sourced diffusion-based model capable of generating complete songs up to several minutes long.

Use cases

• generate a full-length song from lyrics
• create instrumental music from a text prompt
• generate songs in a specific style like jazz or indie folk
• continue or edit an existing song with AI
• run a fast open-source music generation model locally

When to choose

• you need open-source, full-length song generation with a diffusion model
• you want fast inference for complete songs with vocals or instrumentals
• you want to self-host music generation on GPU or macOS

When to avoid

• you need fine-grained DAW-style music editing rather than generative output
• you have no GPU and need real-time generation
• you need licensed commercial stock music with clear rights guarantees

Facets

library · maturity active

audio-processing machine-learning deep-learning artificial-intelligence media python cross-platform music-generation latent-diffusion text-to-music song-generation tts-adjacent audio gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="ASLP-lab/DiffRhythm")