Discussions
Critically Slow Avatar Initialization (10-60 sec) for v1/streaming.new Sequential Calls
Issue Description:
We're experiencing unacceptably long processing times (10-60 seconds) for v1/streaming.new API calls during sequential avatar initialization. The issue is consistently reproducible with 10+ consecutive calls.
Streaming Avatar getting disconnect after 10 to 15 seconds
so below i will be attaching the process of how i am starting the things and let me know if there is any problem in the approach because my session starts correctly and working completely fine but then disconnects suddently after 10 or so seconds and i see in the disconnect room section console log a 5 text only.
Avatar IV via API?
When will be Avatar IV via API available?
Need a Way to Cancel streaming.new if User Closes Window Before SessionID Returns
In our interactive avatar integration, we initiate a session using two API calls: streaming.new followed by streaming.start. However, if the user clicks "Start" and then immediately closes the window, the streaming.new call remains pending (for 20–23 seconds), and we never receive a session_id.
Interactive avatars with custom backgrounds
Hello, is it possible with the interactive avatar API, to use public interactive avatar with a custom background ?
Is it possible to host a HeyGen Interactive Avatar on AWS?
We're trying to deploy an avatar app on an AWS EC2 instance, and while it runs perfectly fine in our local environment, we're encountering an issue in production.
Using personal LLM/API to voice answers with avatar
Hello.
Stream voice to Heygen for speaking back the same voice
I am trying to integrate Heygen with Ultravox. I am sending the voice to Ultravox which gives me the response in voice. How can I stream this voice to hygen for it to speak with facial expressions and lip sync. I couldnt find any good documentation on streaming voice to Heygen.
I tried to use the speak function which accepts text with REPEAT parameter to speak out the text sent. The issue is to pass the entire text returned for LLM take time and the users experience a considerable lag and if I combine few words and send to speak function , the communication doesnt sound natural.
There should be a better way to either stream voice or text to Heygen .
Automatic recording of a streaming session
Since LiveKit supports Egress, I wanted to ask if you provide a built-in option to automatically record streaming sessions. Is this feature currently available but undocumented?
Trouble using "image_asset_id" and external "url" for video backgrounds
Hi,