Discussions

My name is Laxmipriya Kar, and I’m currently working on a Python-based project using Streamlit that functions as an AI Interview Agent. In this setup, the AI takes on the role of an interviewer, dynamically generating follow-up questions based on the candidate’s responses in real time.

The project is functioning well in terms of logic and interaction. However, to enhance realism and engagement, I would like to integrate a talking avatar that can deliver the interview questions as a human interviewer would.

I’ve reviewed your documentation and understand that live avatar interaction typically requires a predefined knowledge base. However, in my case, the conversation is dynamic and driven by live inputs, making a static knowledge base impractical.

Could you please guide me on the best approach to integrate HeyGen's avatar capabilities in such a real-time, unscripted interaction? Are there any APIs, SDKs, or custom solutions you offer that would support this kind of live conversational flow?

Any documentation, examples, or advice you could share would be greatly appreciated.