Discussions

I'm trying to make the avatar listen to the user but not respond automatically. I'd like it to speak (repeat back) my custom generated response. How can I achieve this?

I used startVoiceChat() but then the avatar responds back automatically (I guess with its internally generated response). It's not clear what difference the useSilencePrompt flag makes.

When I used startListening() only instead of using startVoiceChat() it wasn't actually listening, so I wasn't able to capture the user's messages.

Sometimes when I send a speak() request I get the error {"code":400006,"message":"Session is not in correct state"}. What does this mean?

I see that multiple users are facing this problem and there doesn't appear to be a clear solution. The documentation doesn't say much either. Please help!

https://docs.heygen.com/discuss/66f2304586e7d6000f973f5b

https://docs.heygen.com/discuss/66fb057e4c77620020e3bd24

https://docs.heygen.com/discuss/66f5b3cc9507a2001fa54020

Here's the code for how I'm starting the avatar:

const startStreamingChat = useCallback(async (token: string) => {
    streamingAvatarRef.current = new StreamingAvatar({
      token,
    });

    // Some event handling code omitted

    await streamingAvatarRef.current!.createStartAvatar({
      quality: AvatarQuality.Low,
      avatarName: 'Anna_public_3_20240108',
      language: 'en',
    });

    // await streamingAvatarRef.current!.startVoiceChat({
    //   useSilencePrompt: true,
    // });
    await streamingAvatarRef.current!.startListening();
  }, []);

In a separate method I'm using speak() like:

streamingAvatarRef.current!.speak({
  text: customResponse.text,
  taskType: TaskType.REPEAT,
});