Products & Toolsgooglegeminigoogle flowavatars

Google Rolls Out Gemini Avatar and Flow Android

||By LDS Team
6.7
Relevance Score
Google Rolls Out Gemini Avatar and Flow Android
Photo: nokiapoweruser.com · rights & takedowns

NokiaPowerUser reports that at Google I/O 2026 Google introduced a new "Avatar" (internally referenced as "Likeness") entry in the Gemini web interface and a major update to Google Flow with a native Android app and expanded character and scene controls. NokiaPowerUser reports the Gemini web "Avatar" menu is undergoing a gradual rollout but is not yet functional, with the site describing a server-side activation pending. The report describes a workflow where users can type @me to inject a stored likeness into image-generation prompts. NokiaPowerUser also reports Flow's Android debut includes advanced character and scene-building features for creators.

What happened

NokiaPowerUser reports that at Google I/O 2026 Google added an "Avatar" (coded as "Likeness") entry to the Gemini web settings and is rolling that menu out globally, although the feature is not yet active while server-side activation is prepared. NokiaPowerUser reports the site-level code and menu appear live, and the publication describes a user workflow where a stored 3D likeness enables users to insert their image into generated outputs by typing @me into prompts. NokiaPowerUser also reports a major update to Google Flow that includes a native Android app and new character- and scene-building controls aimed at creators.

Technical details

Editorial analysis - technical context

Persistent likeness systems like the described "Likeness" typically require a stable asset store, a rendering or conditioning pipeline, and generation-time identity tokens or shorthand (here @me) to reapply user-specific parameters across model calls. For practitioners, integrating a persistent avatar implies additional service components: secure media storage, identity-to-asset mappings, and runtime conditioning hooks that convert the shorthand token into model prompt conditioning or image priors.

Context and significance

Public reporting places these updates within a broader trend of major AI platforms adding persistent, user-specific media artifacts to generative toolchains to speed workflows for creators. The convenience of @me-style shorthand reduces prompt engineering friction, while a native Android Flow app expands access for mobile-first content production. These moves align with other vendor efforts to bridge model UX and asset management for creators.

What to watch

observers should look for:

  • official rollout status and availability notices from Google
  • any published privacy and data-retention details for stored likeness assets
  • documentation on how @me is implemented (prompt conditioning, tokenization, or server-side merging)
  • parity between the web Gemini experience and the new Flow Android feature set. NokiaPowerUser has not quoted Google directly on rationale or timelines in its report

Key Points

  • 1Gemini web surfaces a new Avatar/Likeness menu, reported as rolling out but inactive, which aims to persist user likenesses for reuse.
  • 2NokiaPowerUser describes a @me workflow to inject a stored likeness into image prompts, reducing repeated reference-image uploads.
  • 3Industry pattern: persistent avatars shift complexity to asset management and privacy controls, important for creators and platform engineers.

Scoring Rationale

Product-level enhancements at a major platform affect creators and developer UX but do not introduce a new model or core research advance. The changes matter for production workflows and asset management, making this a notable product update for practitioners.

Sources

Public references used for this report.

1 source

Practice with real Ad Tech data

90 SQL & Python problems · 15 industry datasets

250 free problems · No credit card

See all Ad Tech problems