zamzam
11/14/2024, 5:26 PMVapi
11/14/2024, 5:29 PMVapi
11/14/2024, 5:29 PMlanguage=multi
parameter in your API request, as you've mentioned. This will enable the system to transcribe audio that contains multiple languages or switches between them. If you're using Deepgram as your LLM provider, make sure to specify the Nova-2 model in your request to ensure compatibility with the multilingual feature. If you need further assistance or have any more questions, feel free to ask!Vapi
11/14/2024, 5:29 PMzamzam
11/14/2024, 5:31 PMShubham Bajaj
11/15/2024, 11:48 AMcodeSwitchingEnabled
to true and then it will work the way your expecting it to use.
After enabling it, you have to instruct the assistant to generate response in user language and use 11labs mutliingual voice which supports possible language user can spoke in.
If you have any more questions do let me know.clumsy.ninja
11/17/2024, 12:11 PMclumsy.ninja
11/17/2024, 12:11 PMclumsy.ninja
11/17/2024, 12:11 PMclumsy.ninja
11/17/2024, 12:22 PMclumsy.ninja
11/17/2024, 12:35 PMShubham Bajaj
11/18/2024, 5:30 PM