In an ambitious move that could change the way we interact with image and video generation tools, Google is working on a new and innovative feature for its artificial intelligence assistant, Gemini. This new feature aims to allow users to create a three-dimensional digital representation of themselves and insert it directly into AI-generated images and videos using very simple text prompts.
According to a comprehensive APK teardown and analysis published by the tech publication Android Authority on March 24, 2026, the feature was discovered in version 17.11.54 of the Google app by well-known developer Assemble Debug. This step marks a notable development building upon a previous feature known as “Personas,” which was first spotted last February in an older beta build.
How does the avatar feature work in Gemini?
The mechanism of the new feature relies on the user scanning their face by recording a short video using their smartphone camera. Afterward, the system processes this data to create a high-resolution 3D model representing the user. Once this step is successfully completed, the user can call upon this model and insert it into any text prompt directed at Gemini to generate images or videos.
To simplify the experience, the personal model is simply invoked by typing the “@me” command or the “@” symbol followed by the user’s username within the chat box. One of the string sequences found within the app’s codebase clearly points to this functionality, stating: “Add yourself to any generated frame by typing @me in your prompt.” The feature appears as a new option named “Avatar” accessible via the assistant’s user interface, supported by a welcome screen that guides users step-by-step through the video capture process.
From mixed reality to standard smartphones
Interestingly, this advanced technology stems from the “Android XR” mixed reality operating system, specifically from the “Avatar” simulation feature launched in late 2025. That feature allowed smart glasses users to create digital avatars of themselves for use in video calls. The system required users to scan their faces using a companion smartphone app, allowing the glasses’ sensors to animate the virtual avatar in real-time to accurately reflect facial expressions.
It now appears that Google is adapting this advanced technology for mainstream use through standard devices, completely eliminating the need to own mixed reality glasses. The previous discovery of the “Personas” list in version 17.4.66 of the Google app back in February hinted at this strategic direction, even though the recording feature was not operational at the time.
The future of visual personalization in AI
The latest discovery in March indicates that the feature has made significant progress in development stages, now bearing a simpler identity along with seamless new integration into content generation workflows. Despite these promising leaks, it should be noted that the feature is not yet publicly available. As with all discoveries resulting from the analysis of incomplete app code, there is no definitive guarantee that the feature will be launched in its current form or launched at all.
To date, Google has not issued any official announcement regarding this feature, nor has it provided a timeline indicating its potential rollout to users. However, if it actually arrives on phones, it will represent a massive and remarkable expansion in personalization capabilities. This move would propel the assistant far beyond merely providing textual context, stepping firmly into the realm of visual identity and giving users a golden opportunity to star in their own AI-generated content without the hassle of uploading a profile picture every time they want to create new material.
The ability to integrate oneself into limitless digital worlds using artificial intelligence opens wide doors for creativity and adds an unprecedented layer of personal interaction, where anyone can see themselves in fantastical scenarios or generate professional profile pictures with the click of a button.
Frequently Asked Questions
What is the new avatar feature in Gemini?
It is an upcoming feature that Google is developing, allowing users to record a short video of their face to create a 3D digital model that can later be inserted into AI-generated images and videos.
How can I use my image in Gemini through this feature?
Once you finish setting up your facial scan, you will be able to type the “@me” command or type “@” followed by your username within your text prompt, and Gemini will place your digital image inside the generated design.
Is this feature available to users now?
No, the feature is still under internal development and was only discovered through the analysis of the Google app’s code. The company has not announced an official release date for the public yet.
Do I need a mixed reality headset to use it?
No, although the technology is built on previous mixed reality systems, Google is adapting it to work seamlessly across standard Android smartphones without any additional hardware.
