Discussions

I’m currently integrating HeyGen’s API into a project and I noticed that, on the website, it’s possible to select an avatar, input some text, and instantly preview the audio before generating the full video. During this preview, the avatar’s mouth is blurred and there’s no full lipsync rendering, which is great for testing purposes and helps reduce unnecessary processing or cost when adjusting scripts.

However, from reviewing the API documentation and endpoints, it seems that each request to generate speech with an avatar requires a full video render, even if it’s just for testing the audio.

Could you confirm if there is any way, via the API, to replicate that quick avatar + text-to-audio preview functionality that exists on the web platform? Ideally, we’d like to validate the voice output before triggering a full video generation.

If this feature isn’t currently supported via the API, are there any plans to add it or perhaps an alternative workaround?