Loccut

Text to speech

Type, pick a voice, and download a WAV made on this device.

How it works

  1. Kokoro runs in this browser. The first run downloads the voice model.
  2. The voices in this release are English.
  3. eSpeak NG speaks Mandarin, and it is stored with the site.

Questions

Where does the voice model come from?

From the model host, if you set one, otherwise from Hugging Face. It is cached in this browser.

Is eSpeak still available?

Yes. The second button uses eSpeak NG, which is stored with the site.

Does the audio leave this device?

No. Decoding and export happen in this browser.

More image tools