HeyGen adds avatar code, a video benchmark and voice cloning for developers
HeyGen’s September releases include an open-source avatar integration, a motion graphics benchmark built with Kaggle and API access to professional voice cloning.
OpenAI
HeyGen introduced three developer offerings during September: software for conversational avatars, a video evaluation benchmark and voice cloning through its API. The company outlined the releases on October 1.
The avatar integration connects OpenAI’s GPT-Live-1 with HeyGen LiveAvatar and animated HyperFrames visuals. Released September 10, it includes examples for tutoring, poker coaching and customer support.
Code2Video, introduced September 18 with Kaggle, evaluates motion graphics generated from 168 creative briefs. It examines visual execution, pacing and adherence to instructions, addressing weaknesses that working code alone does not reveal. HeyGen’s Professional Voice Clone API requires at least 20 minutes of recordings from a single speaker.
The releases give developers components for building video and voice features into their own products. HeyGen says all three offerings are available, while support for testing developers’ own models against the benchmark judge is planned.
The day's biggest AI news in one short email, every morning. Plus an instant heads-up when a reset or a big release lands.
More on OpenAI
- Oct 1OpenAI-hosted essay argues routine work could determine AI’s economic impactOfficial
- Oct 1Higgsfield showcases music release artwork and videos made through ChatGPTOfficial
- Oct 1Higgsfield demonstrates music release planning with GPT-6.1 SolOfficial
- Oct 1The Den says ChatGPT Work saves its leadership team 10 to 15 hours weeklyOfficial
- Sep 30Codex CLI 0.159.3 is outUpdate
- Oct 2OpenAI outlines model selection and production testing for GPT-6 developersOfficial