DiffRhythm
A research diffusion model with inference code for generating full-length songs from lyrics and style prompts.
These buttons open the developer's own site, repository or store listing in a new tab. wares.gg does not host downloads.
About DiffRhythm
DiffRhythm is a song generation model from the ASLP lab. It uses latent diffusion to produce complete songs end to end, and the paper's authors say it is fast and simple to use. The GitHub repository contains the model code, inference and training scripts, a Docker setup and a hosted Hugging Face demo.
The project has continued with DiffRhythm 2 and DiffRhythm+. It suits researchers and technically minded musicians who are comfortable running Python models themselves rather than using a polished app.
Strengths
- Generates full-length songs rather than short clips
- Inference and training code published together
- Docker setup included
- Follow-up versions DiffRhythm 2 and DiffRhythm+ exist
Limitations
- Research code with no graphical application
- Needs a Python environment and suitable hardware
- License terms should be checked in the repository before commercial use
Details
- Pricing
- FreeFree to download and run; the code and model are published on GitHub.
- License
- Proprietary
- Developer
- ASLP-lab
- Platforms
- Linux, Command line
- How it runs
- Downloadable app
- Account
- Not required
- Best suited for
- Researchers and tinkerers who want to run song generation locally
- Categories
- Local AI tools, AI music
- Last verified
- Added
Alternatives to DiffRhythm
Compare allSoftware that can replace DiffRhythm for an important use case, and what changes if you switch.
YuE
Open music generation model that creates songs with vocals, with symbolic planning, covers and editing.
YuE is an open music generation model run from the command line that creates songs with vocals and adds zero-shot covers and editing, with a separate model licence to check.
ACE-Step
Open source local music generation model that runs on your own hardware.
ACE-Step is an MIT-licensed local model running on Windows, macOS and Linux across Mac, AMD, Intel and CUDA hardware, with a clearer open-source license.
Suno
A web-based AI music generator that turns a text prompt or your own lyrics into full songs.
Suno turns prompts or your own lyrics into full songs through a web service with no setup, but runs only online with a capped free tier.
Udio
An online AI music generator for creating, discovering and sharing music tracks.
Udio generates tracks in the browser without a Python environment or local hardware, trading local control for a hosted online service.
Mureka
An online AI music generator that creates full songs, vocals, lyrics and instrumentals from text prompts.
Mureka generates full songs with vocals from prompts or lyrics online, adding music video generation, but uses credits that limit free generation.
Treblo
A free AI song generator, formerly Sonauto, that creates full songs with lyrics and vocals.
Treblo generates full songs with lyrics and vocals free on web and mobile apps, running on the company's servers instead of your own hardware.
Similar software
Related functionality, not necessarily a direct replacement.
Magenta Studio
A free Ableton Live plugin that uses Magenta machine learning models to generate and transform MIDI clips.
Stable Audio
Stability AI's generator for music and sound effects created from text prompts.
AIVA
An AI music composition assistant that generates editable songs in more than 250 styles.
RVC WebUI
A local web interface for training and using retrieval-based voice conversion models.
