View only. Sign in to run this demo.
Text-to-Speech
Convert text into downloadable speech audio using OpenAI or Azure OpenAI.
How to use Text-to-Speech
- 1Write or load a textUp to 3,000 characters. A longer text is cut at the ends of sentences and read in parts, then joined into one audio file.
- 2Choose one voice, or comparePick one voice, or up to three to hear the same text in each of them.
- 3GenerateEach voice gets its own take, with a player, its length and a download.
- 4Listen and keep what you likeTakes pile up on this page: play them one after another, download the ones you want, remove the rest.
- The speech is made by one model through the GAIK TextToSpeech component. The language of the text is recognised from the text itself, and the voice sets how it sounds.
- Takes live in this browser tab only. Reloading or closing the tab loses them, so download the ones you want.
Start from an example
1. The text
Up to 3,000 characters. The server reads 1,000 at a time, so a longer text goes in parts.
0 / 3,000 characters0 words
2. The voice
Pick one voice, or up to 3 to compare them on the same text.
The descriptions are rough guidance: listen to find the voice that suits your text.
Tells the component which language to expect. The speech model also recognises the language from the text.
No audio generated yet
Write a text or load the example, choose a voice (or three to compare), then generate speech. The takes appear here.
