we are currently working on our AI Avatar using eleven labs synth voice. We have a 99USD plan for elevenlabs and a 24USD plan for convai - but it feels like convai uses the custom voice with the cheaper AI voice model “Eleven Flash 2.5” instead of “ElevenLabs Multilingual v2.0”, which causes the voice to sound very bad. Is this assumtion correct and is there a way to get the good quality AI voice model running?
In addition, is there a way to pass voice parameters to eleven labs voice creation (such as speed, stability, similarity, style exaggeration, speaker boost)?
Hi @info_AM , yes we do use “Eleven Flash 2.5” model as default.
But please do provide the config ( parameters ) you want to use with your voice along with the voice name and we can update it on our backend.
I dont know why, but the voice feels slower - please change the speed setting to 1.15. It would realy help to have a user based input for these settings.