AI Text to Speech

How can we help?

Search the documentation or ask the AI agent anything about the plugin.

Use more than one voice in a post

Updated September 15, 2026

With PRO, a single audio file can use several voices. Wrap any part of your post in [voice="name"]...[/voice] and that section is spoken in the named voice. Everything outside the shortcodes uses the main voice you pick in the AI TTS box. This works well for quotes, interviews, Q&A posts and short stories with dialogue.

Add a second voice

  1. In the post, type [voice="shimmer"] just before the text that should use a different voice.
  2. Type [/voice] where the main voice should take over again.
  3. Save the post, pick the main voice in the AI TTS box and click Generate TTS.
I asked Megan which CMS she prefers. [voice="shimmer"]It has to be WordPress. There's a plugin for almost everything.[/voice] Her answer didn't surprise me.

[voice="onyx"]Question two: how long have you been building sites?[/voice]
[voice="nova"]About eight years now.[/voice]

The shortcodes are hidden on the published post, so readers only see the text.

Voice names

Use one of these, in lowercase: marin, cedar, alloy, ash, ballad, coral, echo, fable, nova, onyx, sage, shimmer, verse.

If the name isn’t recognised (for example a typo), that section falls back to the main voice without an error. If you use the TTS-1 model, choose voices that model supports; see Voices, voice models and supported languages.

How it behaves

  • The spoken title, author and date intro always uses the main voice.
  • [/voice] returns to the main voice, not to the previous voice. Voice shortcodes don’t nest, so close one before opening the next.
  • The selected style and any custom voice instructions apply to every voice in the file.
  • Each voice section is sent to OpenAI as a separate request and the parts are joined into one MP3. Many short switches mean more requests and a slightly longer generation time. The characters are billed the same.
  • Because sections are generated separately, there can be a small change in volume or pacing at each switch.

Use straight quotes around the voice name. Curly quotes, often added when pasting from a word processor, stop the shortcode from working.

Related

Was this article helpful?