Hey, I’ve handled this issue before and can help you fix it. It’s usually caused by how Vapi processes or encodes audio before sending it to the model. Could you share your current Vapi transcription setup or audio format? That’ll help me identify the cause and guide you better. @ClaraThays