Scylla’s Band demo
Try it in your browser
Type something and hear it in any of the ten voices. The model runs on your own computer, inside this page.
Before you start
- It’s a 261 MB download. Your browser fetches the model from Hugging Face once and keeps it, so later visits start without downloading. You can delete it at the bottom of this page.
- It needs about 1 GB of free memory while it runs. Tested in desktop Chrome. Phones and tablets may run out of memory.
- It’s slower than the native runtime. Speech is generated on your CPU in WebAssembly, several times slower than the Python or C++ runtime. On a recent desktop it still runs ahead of playback, so audio starts within a few seconds.
- Nothing you type leaves your computer. The only network traffic is the model download.
Speak
0 of 600 characters
Stored model
Nothing is stored yet.