Natural-sounding text to speech without a lot of setup
A look under the hood of the Text to Speech tool.
5 MIN READ
The Text to Speech tool wraps a managed neural speech synthesis engine. On submit, the tool page posts your text, a chosen voice, and an optional speaking style to `/api/synthesize`, which builds the SSML payload server-side and streams back an MP3.
Styles like `cheerful` or `empathetic` only work on a subset of neural voices, which is why the voice picker filters styles based on what the selected voice actually supports — better to grey out an option than let a request fail silently.
Because synthesis happens server-side, your API key never touches the browser, and the audio is generated fresh on every request rather than cached anywhere.
Try the tools this post is about
Free, no account needed.