Discussions

We are currently integrating HeyGen avatars into our platform and would like to explore advanced API capabilities. Specifically, we are looking for a way to send pre-recorded audio directly to the avatar instead of relying solely on text input for speech generation.
The reason behind this request is that the current avatar voice synthesis does not handle Moroccan Darija or Arabic fluently. As a result, when we send Arabic or Darija text, the pronunciation and delivery are not accurate or natural, which affects the user experience.
We would love to know:
Is it possible to send audio input directly via the API for the avatar to lip-sync?
If not, is there a roadmap or plan to support more interactive languages such as Arabic or Darija?
Can we collaborate or request support to help enable these features?
Looking forward to your response and hoping we can find a way to enhance the experience for Arabic-speaking users.