Could minimax voice add

pronunciation_dict

parameter to define pronunciation rules for specific characters or symbols?

Hi Ben, Minimax doesn't support ElevenLabs-style pronunciation dictionaries. For pronunciation control use MinimaxVoice.languageBoost and textNormalizationEnabled (see changelog: https://docs.vapi.ai/changelog/2025/9/11). There is no built-in

pronunciation_dict

parameter in the MinimaxVoice spec—only language hints, normalization, pitch/speed, and other voice parameters shown in the API schema (see voice fields in Sessions API: https://docs.vapi.ai/api-reference/sessions/update). If you need per-word phoneme/alias control today, use an ElevenLabs voice with a pronunciation dictionary (docs: https://docs.vapi.ai/assistants/pronunciation-dictionaries). I can file a feature request for adding a

pronunciation_dict

-like parameter to Minimax—want me to do that?

Hi Aniah, Thank you for your reply. I just checked the Minimax newest API document, I found that the API interface already supports the

pronunciation_dict

parameter, please refer to https://platform.minimax.io/docs/api-reference/speech-t2a-http#body-pronunciation-dict I hope this feature will be supported soon. Thanks.

Thanks for getting back to us. We are delighted to know your issue has been resolved. Please do let us know if you need further assistance. We will be closing this for now and have a great day ahead!