NVIDIA has made its Audio2Face AI technology open source, enabling developers and researchers to create lifelike, talking avatars and drive innovation in animation and gaming.
NVIDIA has taken a significant step in democratising AI-driven animation by making its Audio2Face technology open source. The move allows developers, students, and researchers to access the models, SDK, and training framework, enabling them to fine-tune it for custom characters and applications.
Audio2Face transforms audio phonemes and intonation into animation data, which is then mapped onto a character’s face to produce life-like, talking avatars. The technology can be used offline for pre-scripted scenes or streamed live for dynamic, AI-driven interactions.
Several studios and developers are already leveraging Audio2Face to enhance production and realism. Notable users include Convai, NetEase, Codemasters, GSC Games World, Inworld AI, Perfect World Games, Streamlabs, and UneeQ Digital Humans. Platforms like Reallusion have integrated it into their 3D character creation tools, while game developers Survios and The Farm 51 use it to accelerate workflows and push animation fidelity.
To foster a broader AI animation community, NVIDIA has also launched a Discord forum for developers to share work and collaborate. The company expects this open source release to fuel the growth of lifelike AI avatars across games, media, and other interactive content.
Highlighting the potential, NVIDIA notes:
“With this move, it looks like lifelike, AI-powered avatars will be everywhere and anywhere when there’s a need for NPCs to talk in games, media, or beyond.”
This initiative reflects a growing trend among major AI innovators to support industry growth via open access, helping democratise advanced tools and accelerate innovation across animation and interactive media.
