With Gemini Avatar features, you could soon insert a digital version of yourself into AI-generated images and videos without having to pick up your camera. A recently created Google app hints at tools that will scan your face and turn it into a reusable 3D model in Gemini, as seen in Android Authority’s APK breakdown.
The concept is based on Android XR’s previous Likeness system, which created a realistic replacement for video calls. Now it goes beyond headsets. The same approach is appearing on Gemini, with signs that it could work on phones and desktop browsers.
Instead of taking a new selfie every time, you would create a permanent version of yourself once and reuse it for different editions. This shift favors speed, but also helps keep your appearance consistent if you create content frequently.
Apple already offers Memoji, a stylized version of you for Messages and FaceTime. Google’s approach seems more realistic and is designed to work directly in generative tools and not just serve as a communication function.
This is how the Gemini avatar works
The system appears to be based on a straightforward recording process. You record a short video of your face and the app converts it into a 3D model. It follows the same basic idea as Likeness, but does not require a headset.
On-screen instructions guide you to frame your face and capture enough detail. Once the model is complete, it will be saved to your account and linked to your identity across all Gemini features.
A notable detail is how the setup works. The creation flow appears to be web-based, which suggests you may not be limited to your phone. A desktop option would simplify the process and emphasize that this is intended to work across devices.
More than Apple’s Memoji
Memoji focuses on expression and simplicity, using a cartoon style that suits messaging and video calls. Google’s approach aims for a more realistic result that blends in with the generated images.
Instead of sending an animated face, you could place a lifelike version of yourself into AI-generated scenes. The build hints at prompts that let you insert yourself directly into content and hints at deeper integration with Gemini’s creative tools.
This changes the way you create. Instead of taking new images every time, you work with a reusable asset that can be inserted into different scenarios.
There is a compromise. A more realistic scan raises questions about accuracy and privacy, and Google hasn’t shared details about how it will address these concerns.
What awaits you next
The feature is still under development and has not yet been released in a functional form. Since this is an app teardown, the final version may change or not launch at all.
Nevertheless, the direction is becoming clearer. The change from previous names like “Character” to “Avatar” suggests a broader identity system associated with Gemini. Combined with cross-device support, this looks like more than an XR experiment.
If released, it will likely show up first in Gemini’s creative tools, where it would be most valuable for quickly blending into visuals. There’s no timeline or confirmed availability yet, so for now this remains a first look at where Google’s AI tools are headed.
