check out the modell-settings. in my case the model ( Voice ) was not set. You can choose between eleven_multilingual_v2 if transcriberlanguage is non-English, otherwise eleven_turbo_v2.