Discussions

How can we integrate custom LLM like open AI for voice chat with Avatar, currently whenever asked it always replies with its knowledge first, which we don't want. It should only speak according to the speak method of streaming avatar sdk or use knowledge of custom LLM.

Hi, there

We have a huge amount of information that we use for training our agents. How much knowledge can I add to the base knowledge? Do I need to do something else to make this work?

Good afternoon, I want to generate several hundred videos via API with one video avatar, which I will create in advance in the web interface. These videos will be differentiated only by the name that the video avatar pronounces.

this is the full request

Overall scene duration exceeds overall script duration.
The scene exceeding the script duration will not have TTS (text-to-speech) audio. If there is an Avatar in this part of the scene, it will not speak.

I'm trying to connect my interactive avatar with my GPT from OpenAi.
I keep getting this error: SyntaxError: The requested module '@heygen/streaming-avatar' does not provide an export named 'StreamingAvatar'. I tried using 'default' and that did not work either. What do you recommend?

What is being used for transcripting the user voice in Hey Gen Streaming Avatar SDK?

My interactive avatar has stopped talking to me. It can hear me and all conversation is appearing on the screen including her greeting and prompts for me to continue engaging. She remains in the idle mode not frozen but not speaking.