Guide · Live avatar
Gemini 3.8 Live with Live Avatar: access, languages, and limits
Google announced Gemini 3.8 Live with Live Avatar on September 24, 2026. It adds a responsive visual persona to a live spoken conversation, with tool calls that can run while the agent keeps talking. It is an enterprise product for interactive agents. It should not be confused with the personal avatar feature in Gemini Apps, which is used to create images or videos from a user's likeness. The two features have different access rules.
Where is Live Avatar available?#
Google's announcement says Gemini 3.8 Live with Live Avatar is available in Gemini Enterprise. A separate Google Cloud announcement names US and EU endpoints with provisioned throughput, enterprise compliance, and data governance. Neither of those announcements provides a country-by-country end-user availability list. For a deployment, verify the current endpoint, account entitlement, data residency, and enterprise terms with Google before promising access in a particular country.
The Google Cloud article calls the feature generally available. That status does not mean every customer has every customization option. Google's launch article states that custom avatar creation from a reference image requires enterprise allowlisting. A library of preset avatars is described separately. These details were checked on 2026-09-27 from the Google launch article (external site) and Google Cloud availability article (external site).
How many languages does it support?#
Google says Live Avatar can transition across 97 languages while adapting lip sync and expressions. The claim applies to Gemini 3.8 Live with Live Avatar in Google's announcement, not to every Gemini Apps avatar workflow. The official page does not supply a per-language quality guarantee or a complete deployment-region list. A multilingual pilot should test the exact languages and code switching used by its audience, including names and domain vocabulary. Source: Google announcement (external site), checked 2026-09-27.
What are the known restrictions?#
A custom visual identity based on a reference image requires enterprise allowlisting. The Google announcement says generated audio and video are watermarked with SynthID. The feature is presented for Gemini Enterprise rather than as a general, anonymous web tool. Google describes asynchronous tool execution, but that does not establish latency or reliability for a particular production integration. Check the current API documentation and contract before making a service-level promise.
The name “Gemini avatar” also appears in Gemini Apps Help. That separate personal avatar feature requires a user aged 18 or older, signed in with a personal Google Account; Google's help page says it is currently unavailable in the EEA, Switzerland, and the United Kingdom, and supports English only. Creating a video with that personal avatar requires a Google AI plan. These restrictions are about the Gemini Apps personal avatar, not the 97-language enterprise Live Avatar feature. Source: Gemini Apps Help (external site), checked 2026-09-27.
Is this a free AI video generator?#
No free tier or unlimited use is asserted here. Live Avatar is positioned as a live enterprise conversation capability. If you need a finished video file, start with the text-to-video guide or image-to-video guide. If you need an interactive agent, use the live avatar scenario comparison to test support, commerce, and education requirements before selecting a vendor.
Sources#
Official pages checked 2026-09-27: