Foundational AI avatar model
K8
A llama3.2 11B based multi-modal AI that accepts text, video and audio input and generates a real-time AI avatar output.
- Image in, real-time AI avatar out.
- A state-of-the-art diffusion transformer (DiT) model generates real-time synthetic video and audio, synced to an input script and image.
- Low latency, low bandwidth.
- A unique encoder-decoder compression algorithm enables real-time video and audio without the transmission of binary video or audio data.
- Saftey built in.
- Invisible watermarking and a unique AI model that detects and blocks the generation of harmful content ensures that the AI avatar is safe to use in any environment.
AI assisted education
eduk8
An AI native online education platform that allows students to learn at their own pace and participate in AI tutor led collaborative learning sessions.
- Dynamic learning.
- Generate rich learning experiences from images, video or interactive content. Discover how topics relate, and build a dynamic learning path that is unique to you.
- AI tutor.
- Discuss ideas and ask questions in real-time with your personal AI tutor, available via text or video chat.
- Collaborative learning.
- Join others on a similar learning path and participate in scheduled collaborative learning sessions assisted by a group AI tutor.