Discussions

voiceChatTransport cuts Interactive Avatar off and injects messages we didn't specify

We’ve been trying to track down the cause of increased latency for our non-US users. We went as far as setting up a new environment in another region for our application but found that the bottleneck seemed to be the avatar. While troubleshooting internally, we toggled the voiceChatTransport to “livekit” via the SDK’s createStartAvatar function and that seemed to immediately help significantly with latency, but the very first message we tell the avatar to repeat after the connection is established gets cut off and it starts speaking a different message we didn’t send. Is there a way we can have voiceChatTransport set to "livekit" without that undesirable message cutting off the text we asked the avatar to repeat?

1 hour ago by Scott L

Answered

[ISSUE] matting: true causes HTTP 500 Internal Error on talking_photo (API v2)

Summary

3 hours ago by NgB

Answered

Emotions via API

Hello Team!
I love your tool and how AI generates shorts with avatars. However I can not conceive how can I send emotions (pronunciation) generated with the AI like:

11 hours ago by Oleksandr

Answered

500 errors from Avatar IV API for new image uploads

I uploaded a jpeg image through the Upload Asset API and received a 100 response with an id, name and image_key etc. that all look fine.

I waited 30 minutes to let it process on the HeyGen backend and verified it was showing up with the List Assets API

I then tried to use that image key in an Avatar IV generate request (https://api.heygen.com/v2/video/av4/generate) and received a 500 internal server error:
{
"data": null,
"error": {
"code": "internal_error",
"message": "An unexpected error occurred"
}
}
I retried the exact same request but used an image_key that I uploaded a few weeks ago (no other changes) and the request works fine and successfully generates a video. This was the exact same jpeg file as the one that is now giving errors.

21 hours ago by BG

Answered

Lip-Sync

I am using a custom avatar that I created with the Azure speech LLM , I am having a difficult time getting the movements of the avatar and the speech to sync up. It looks like I am lip-syncing a song. Can you offer advice?

22 hours ago by Ken