Discussions

Hi team,

We're building a virtual university advisor where maintaining accurate, knowledge-based responses is crucial. Our main goal is to keep the full retrieval functionality but significantly reduce the initial response latency.

Hi, im trying to integrate elevenlabs voice and response to bypass the built-in voice mode response of interactive avatar, how it could be possible? Which function handles this basic response?

Hello,

Is is possible for the LLM to get its knowledge from a folder that contains text documents rather than the URL of a website ? or can the URL address be the folder location ?

We have integrated API using Streaming API v2 Integration with LiveKit since we need to use it in Sawtooth and require raw files. Here’s the reference link:
Streaming API Integration with LiveKit v2
Below are two questions can you please response we can plan next step accordingly
1- Can we use replace-text-in-template with the Streaming API v2 Integration (LiveKit), as we have implemented HeyGen using this method?
2- If we need to integrate the SDK, can we implement it in JavaScript?

Hello,

Hello!