Discussions

Hi HeyGen team — quick question about voice availability across products.

I would like to create a live avatar but dont know where to start. I have tried to setup using a preset avatar but find the lack of information difficult. I would like to setup in sandbox mode so I can completely develop and test and see if LiveAvatar will fullfil my requirements before commiting my $99

I'm trying to generate an avatar using Liveavatar, which requires a video with specific specifications. I generated the video and the character in it using AI, but after uploading the video, it asks me to confirm that the character in the video is me, even though it's an AI-generated character. Can you guide me on how to overcome this issue and complete the avatar generation process?

Liveavatar (Full Mode) with user avatar is not lip syncing and playing very slowly (low frame rate)
What could be the reason, avatar has chrome key on it.

Hi!
I don't see a language parameter in the documentation to start a convo in a certain language, but am I correct to understand that LiveAvatar cannot switch languages during the convo? When I switch from English to Dutch, the SST seems to 'want' to hear English still.
It's unclear to me what the internal logic is.

Hi. I'm actually backend developer but i need to hologram for my customer. I was deal with demo app but i couldn't get it. How can i use sdk with html elements. I put video element on html side but i couldn't displaying on it. Did you have any html-based project ?

Hi!

While I'm connecting to ws_url sometimes the connection is working fine but in few cases it is connecting and then closes immediately.
Is there something out of these 3 points:
1.Only have frontend connect to the WebSocket 2.Use a different backend control mechanism (REST API, separate WebSocket endpoint) 3.Ensure only one client connects at a time (coordinate frontend/backend)

In HeyGen, I created a system that used a third-party STT, then sent the recognized text to the Streaming API, and the avatar responded through my prompt using knowledge_base. Can I do the same in FULL MODE? Because after reading the documentation, I didn't understand if I could do this.

Does live avatar has a feature of Acknowledgement if the query is taking longer to respond - or can we do it custom, will it affect streaming?