Introducing Gemini 3.8 Live with Live Avatar

Google has launched Gemini 3.8 Live with Live Avatar, integrating real-time video streaming with its conversational AI to create more natural interactions. The feature pairs low-latency video generation with speech, allowing agents to listen, see, and respond with dynamic visual personas. Enterprises can deploy these avatars for customer service or interactive walkthroughs, transforming digital exchanges into richer experiences. The technology is now available in Gemini Enterprise, expanding virtual offerings with expressive, responsive agents.

Live Avatar enhances conversations by processing visual and audio inputs simultaneously, enabling agents to respond with synchronized expressions and lip movements. The feature supports asynchronous tool calls, allowing background data fetching while maintaining uninterrupted dialogue. This capability lets agents handle complex tasks, such as hotel check-ins, without disrupting the conversational flow. The system leverages Gemini’s advanced reasoning to manage multiple inputs seamlessly.

The feature includes native multilingual speech-to-speech synchronization, adapting lip-sync and expressions across 97 languages without degrading video quality. Organizations can customize avatars to align with brand identities, using preset options or generating fully animated avatars from reference images. Custom avatar creation is currently restricted to enterprise allowlisting, ensuring controlled deployment and adherence to brand guidelines.

Google has implemented safeguards to maintain transparency and accountability, including SynthID watermarking for AI-generated audio and video output. This imperceptible watermark helps detect AI-generated content, reducing risks of misinformation and misattribution. The feature is available in Gemini Enterprise, with API documentation provided for integration. Organizations are encouraged to review the model card for safety and responsible deployment guidelines.