Discussions

We are currently researching and testing. I had a few questions:

1. Is there a limit to the amount of KBs we can have in our KB library; for our usecase, we will need to create 10s and 100s of KBs for each of our client. I want to know if there is an upper limit that I should be aware of.

2. Is it possible to receive preview thumbnail url for interactive avatars, as we do for video gen avatars, when using list all avatars api endpoint. It is a big necessity to scale the product we are building. We plan on using a different matching certain physical cues like age, gender, etc. for that we need the preview thumbnails.

3. Is it possible to get streaming quality above 720p for enterprise users?

4. I have noticed that at times there is lag between lipsync and sound, even though lag is ~1 second, it feels very weird. Is there a way to mitigate that.

5. Is there a way to receive internal transcript that is converted to sound via eleven labs maybe? That internal transcript will be beneficial to refine our LLMs that will generate the KBs

6. For enterprise users can we discuss something like a lumpsum payment for interactive avatar slots as opposed to current recurrent monthly price. We are looking at having ~100-200 interactive avatars in our library as we scale.

7. I have noticed an bug. If I create a new KB via HeyGen labs UI or API endpoint, and set an opening intro; when I use the KB with a interactive avatar, it never speaks out the intro. Maybe its because I am streaming over on my lite UI which I built using LiveKit's CDN. Though I want to know what's happening here. Cause once the session is created it is started within the same API call, and time taken is almost ~1.5 seconds. I think it shouldn't be done with opening into within that time