AI Music GeneratorBeta

Describe a piece of music and an open AI model writes a short instrumental clip on your own device — saved as a WAV file.

Describe a piece of music and an AI model writes it — a short instrumental clip, generated on your own device and saved as a WAV file.

You need a recent browser, some free disk space, and patience on the first download. A GPU (WebGPU) makes everything many times faster and is used automatically when this device passes a real GPU test; without one, everything runs on the CPU.

Choose a model and press Download & load to begin.

The first load downloads the weights into this browser's storage. Later visits reuse them and start in seconds.

Resource usage no model loaded
CPU load 0% busy
CPU threads
Memory
GPU
Model storage
Output speed

Browsers do not report CPU or GPU load to a web page, so nothing here is invented. CPU load is measured two ways and shows the higher: how much of each second the engine spends blocking its own worker, and how much slower a fixed piece of arithmetic runs right now than on an idle machine — which is what a multi-threaded engine looks like from here. Memory is the JavaScript and WebAssembly heap, which only Chromium-based browsers expose. GPU utilisation and video memory are not available in any browser.

Generation settings

Create an image

No model loaded yet.

Speech to text, text to speech, music and images share one model slot, so loading one replaces the previous one — your chat model stays where it is.

Transcribe audio

No model loaded yet.
No audio chosen yet.

Speech to text, text to speech, music and images share one model slot, so loading one replaces the previous one — your chat model stays where it is.

Read text aloud

No model loaded yet.

Speech to text, text to speech, music and images share one model slot, so loading one replaces the previous one — your chat model stays where it is.

Make music

No model loaded yet.

MusicGen is slow without a GPU — measured on a plain CPU, about 50 seconds of computing for 2 seconds of music. Start with a short clip.

Speech to text, text to speech, music and images share one model slot, so loading one replaces the previous one — your chat model stays where it is.

Load any model from Hugging Face

Search the Hub, or paste a repository id. Each weight variant is listed with its real download size and whether it is likely to fit on this device.

Downloaded on this device

Models are stored by your browser and count towards its storage quota. Delete any of them here at any time.

Saved conversations

Your chats are kept in this browser only, so you can pick one up on your next visit. They are never uploaded, and deleting models above does not touch them.

This device

Graphics drivers and browsers get updated. If GPU inference was switched off after a failure, you can ask for it to be tested again.

Log


  

🔒 Runs entirely in your browser — nothing you type is uploaded or stored on a server.

The AI Music Generator writes a short instrumental clip from a description, entirely inside your browser. The model is Meta's MusicGen Small: tell it a genre, the instruments and a mood — “lo-fi hip hop with a mellow piano and vinyl crackle” — and it composes the audio from scratch.

Also on Txtset: ask a local AI model for prompt ideas · record a voice-over to go on top · run speech recognition on the same device · draw cover art for the track.

It is a real generative model, not a loop library, which means two runs of the same prompt give you two different pieces. You choose the length (2 to 20 seconds) and how closely it should follow the prompt, and every clip can be saved as a WAV file for a video, a game prototype or a podcast intro.

Be ready to wait. Music generation is heavy work: on a plain CPU we measured about 50 seconds of computing for 2 seconds of music, so start with a short clip. A computer with a working GPU is much faster. The model is about 650 MB and is downloaded once.

How to use

  1. Press Download & load (about 650 MB, once).
  2. Describe the music: genre, instruments, tempo and mood.
  3. Set the length — start short on a CPU — and press Compose.
  4. Listen, then press Download WAV to keep it.

Examples

Video background
A few seconds of calm ambient music under a product clip.
Game jam placeholder
Quick mood music for a prototype level.
Podcast sting
A short intro or transition sound.
Exploring styles
Hear what “80s synthwave” versus “baroque harpsichord” does to the same idea.

Frequently asked questions

Can I use the music commercially?
MusicGen's weights are released by Meta under the CC-BY-NC 4.0 licence, which does NOT allow commercial use. Use the clips for personal projects, prototypes and learning; for a commercial product, use music you have a licence for.
Does it make vocals or lyrics?
No. MusicGen generates instrumental audio only.
Why is it so slow?
Each second of music is 50 audio tokens that the model generates one after another, and a CPU does that slowly. A GPU (WebGPU) speeds it up a lot; on a CPU, ask for short clips.
Is my prompt uploaded?
No. The model runs in this tab; only the one-time model download touches the network.
What is the longest clip?
20 seconds per run here. Longer clips need more memory and a long wait on most devices.