Every audio file starts with a language and a voice. TTSWP comes with hundreds of voices across 70+ languages. Pick your default once. Override it per post or per shortcode when you need to.
Open the Voices page
Go to Text to Speech → Voices. The page shows three things at the top:
- A language picker (left)
- The currently selected voice (middle)
- A refresh button and model selector (right)
Below those, the voice list shows every voice available in your chosen language.
Pick a default language
The default language decides which voices appear in the list and which language TTSWP sends to ElevenLabs for pronunciation. Pick it first, then pick a voice.
- Click the language dropdown.
- Choose from the list.
- The voice list below refreshes automatically.
Note: On multilingual sites using WPML or Weglot, the language picked here is the fallback. Each translated post can use its own voice. See Language voice mapping.
Preview a voice
Before you commit to a voice, listen to it.
- Find a voice in the list.
- Click the Play icon next to the voice name.
- A short sample plays. The sample uses the current language and a pre-set test sentence.
Previews do not use your credits.
Set a default voice
Once you find a voice you like, click Select. A green checkmark shows which voice is your default.
TTSWP uses this voice for every audio file, unless:
- You override it per-post with the voice dropdown in the editor
- You override it in a shortcode with the
voiceattribute - WPML or Weglot is active and uses a different voice per language
Voice categories
Voices come in a few types. The badges next to each voice name tell you what type it is.
- Standard voices: included on every plan
- Neural voices (PRO): the premium natural-sounding voices
- Cloned (PRO): voices you have cloned in your ElevenLabs account and referenced here
See Voice cloning to use your own cloned voice.
Model selector
The model is the ElevenLabs engine that actually produces audio. Different models have different strengths.
| Model | Best for | Per-request limit | Languages |
|---|---|---|---|
| Eleven Flash v2.5 | Fast generation, popular languages | 40,000 chars | 30+ |
| Eleven Turbo v2.5 | Balanced speed and quality | 40,000 chars | 30+ |
| Eleven Multilingual v2 | Natural sound, long-form content (PRO) | 10,000 chars | 30+ |
| Eleven v3 | Most expressive, podcasts and narration (PRO) | 3,000 chars | 30+ |
Rule of thumb: Flash v2.5 is the default for most sites. Multilingual v2 if you want the most natural voice and your posts are under 10,000 characters. v3 for podcast-style or emotional narration.
Refresh the voice list
Click the Refresh voices button in the topbar if:
- You added a new voice in your ElevenLabs account and do not see it here
- The list looks empty or shows a stale selection
Refreshing re-fetches your voice list from ElevenLabs. It does not cost credits.
Change voice on a single post
If you want one specific post to use a different voice, override it in the post editor. The plugin adds a small sidebar on the post editor called TTS audio. Inside that sidebar there is a voice dropdown. Pick the voice you want for that post.
See also the Shortcode Generator if you want custom voices inside shortcodes.
Common questions
Why can I not select certain voices? Neural and cloned voices need a paid plan. On free plans they are shown but locked.
Why is my preferred voice missing? Check the language picker. Voices are filtered by language. Switch to "All languages" to see every voice.
Why does the voice sound robotic? Try a different model. Flash is optimized for speed, Multilingual for naturalness. Also consider Word replacement for specific word pronunciation issues.