How to make the avatar repeat the text which I provide? avatar. Speak() seems to reply the text provide.

I am trying to set up the interactive avatar using the git demo (InteractiveAvatarNextJSDemo). Unfortunately, I can't get it to work even though I followed the instructions. even after adding an openai key to the env file, there is no way to use whisper in the browser, nor can the avatar access chatgpt4. In general, the instructions seem imprecise or incorrect to me. I have generally only been able to interact with the avatar after modifying the code in "packages". In the dev specification, a "node " must be added at the beginning of the string, otherwise an error is generated that the path to "node_modules" does not exist. It would be great if someone who has managed to link the demo to their own LLM could point out where changes are needed.

Has anyone encountered HeyGen dropping the first few words when an Avatar starts speaking? The mouth moves but no sound comes out. I've tried adding spaces, delays, etc. but nothing seems to work.

Hi, I am Wang based in Japan. Hope you are doing well.

The acronym for our association is ISTA. We pronounce it as if it's a word. It's pronounced exactly like the word VISTA but without the "V"... with the emphasis/accent on the first syllable. We use the voice called ECHO for all our HeyGen work. Your system pronounces it correctly every once in a while, but most often, it pronounces it like with a long I (like "eye" and then "sta"). I'm familiar with using substitutes and spelling it phonetically. I've done that a lot with other words and I've worked through it. Unfortunately, that hasn't worked for the name of our company. I've tried "istuh" and others. No luck. It still doesn't work consistently (or at all)... and the voice often starts accenting the second syllable. Any suggestions for how to fix this and lock it in? TIA

I am writing to inquire about the API services offered by HeyGen. Specifically, I would like to understand the functionality and pricing details of the available features.

Hello,