Discussions

Hi, I just created a new custom avatar and it seems to be stuck at 20% ready for more than 30min.

Hola a todos,

I wanted to clarify—does the 4000-character limit apply specifically to the knowledge base entries

In interactive avatars, the avatar's speech can be obtained as a string from the "StreamingEvents.AVATAR_TALKING_MESSAGE" event. However, this string can be obtained before the avatar actually speaks. Is it possible to obtain the string as soon as possible in real time along with the avatar's actual speech？

Is it possible to use an interactive avatar with its own knowledge base directly on HeyGen's website and extract the conversation to an external database?

I'm currently running a custom web app that integrates the Heygen Interactive Avatar SDK (deployed via Vercel), used on an NVIDIA Shield Pro (Android TV). The avatar loads and responds as expected, but I’ve run into a specific issue with Bluetooth audio:

I'm experiencing an issue with the HeyGen API where my video generation requests reach approximately 97% completion before failing. Here's my setup and what I've tried:

Currently, regardless of who creates a knowledge base, it can be accessed by any account, leading to potential information leaks, as the Heygen API is called on the client side.

I am confused why I can’t find any avatars?publicTab=ugc> using the list-avatars-v2> endpoint. Can someone please help?

Hi, I want to use animated/cartoonish Avatars for streaming instead of Avatars that look like real people.