When WPML is active, TTSWP can pick a different voice for each translated language. A Spanish post gets a Spanish voice, an English post gets an English voice, and visitors hear the right language automatically.
Requirements
- WPML 4.0 or newer
- At least two languages configured in WPML
- TTSWP connected to your site
Turn it on
- Go to Text to Speech → Overview.
- Scroll to Quick Settings.
- Flip WPML integration to on.
- Save.
If WPML is not installed, this toggle is locked and shows "WPML not detected".
Map voices to languages
Once WPML integration is on, a new page appears: Text to Speech → Language Voice Mapping. See Language Voice Mapping for the detail.
Each WPML language gets its own voice. Fallback to the default voice if no language-specific voice is set.
How it works at generate time
When you generate audio for a translated post:
- TTSWP looks up the post's WPML language code
- Finds the voice mapped to that language
- Uses that voice for synthesis
Existing audio on translated posts is not automatically regenerated. Regenerate manually after changing voice mappings.
Language code normalization
WordPress uses nb and nn for Norwegian. ElevenLabs uses no. TTSWP normalizes these automatically so Norwegian posts work out of the box.
Known quirks
- If a post is translated but no voice is mapped, the default voice is used (not an error)
- Post types not enabled in Content settings are ignored even if WPML translates them
- WPML String Translation is not used - we only read the language code