Discussions

Hi Team!

We are using the HeyGen streaming endpoint to create a session on the API that generates livekit rooms. Our goal is to integrate HeyGen streaming avatar without deploying additional infrastructure.

We are facing an issue where we are not receiving any transcriptions for either the AI agent or our participant speech. We are currently listening for RoomEvent.TranscriptionReceived from the livekit room

To resolve this issue, in order of priority:

I) Preferred Solution: Can you please enabling STTSegmentsForwarder in the room? See link This solution is straightforward and would make things a ton easier with no overhead for HeyGen team either.

II) Less Ideal: Issuing Multiple LiveKit Tokens: If the first solution is not feasible, a second option would be to allow getting multiple LiveKit tokens for the room created. This would enable us to deploy an agent that can join the LiveKit room as a third participant and perform transcription on the participant audio streams using the livekit SDK.

What can we expect from HeyGen in this regard? We'd love to be able to use the streaming avatars. P.S. The emotive avatars are great but not available for streaming. Any word on the feasiblity of them coming to streaming endpoinits?