Google Photos Prepares Massive ‘Video Remix’ AI Upgrade
Google is testing a “Video Remix” AI tool in Google Photos, codenamed “Soba,” according to forensic app specialist AssembleDebug. The feature likely leverages the Gemini Omni model to transform existing video clips using text or voice commands, moving beyond simple photo-to-video conversions to offer generative AI video editing.
What is the “Soba” feature in Google Photos?
Soba is the internal codename for a forthcoming AI-powered video editing tool discovered within the Google Photos Android app. AssembleDebug uncovered a new button in the app’s “Create” tab featuring a YouTube-style play icon overlaid with Google’s Gemini “sparkle” logo. While the functional code isn’t active yet, the assets indicate a shift toward generative video modification.
Internal test strings show that enabling Soba triggers a name change for an existing tool. The current “Remix” button, used for AI photo edits, rebrands as “Photo Remix” to make room for the new video-centric functionality. This naming convention suggests Google intends to create a parallel ecosystem for both still and moving images.
How does Video Remix differ from existing AI tools?
Video Remix is not a “photo to video” tool. According to reporting by Paul Monckton for Forbes, Google’s current offerings create clips from still images. Soba takes an existing video clip as input and produces a modified version as output.
The technical backbone also differs. While the current Photo Remix feature relies on the Nano Banana model, Soba is expected to use Gemini Omni. This allows for “conversational editing,” where a user could theoretically tell the app to change a home movie into a claymation-style animation using natural language.
| Feature | Input | AI Model | Output |
|---|---|---|---|
| Photo Remix | Still Image | Nano Banana | Edited Photo |
| Video Remix (Soba) | Video Clip | Gemini Omni | Modified Video |
Will users get full conversational AI video editing?
It’s unlikely that the full power of Gemini Omni will arrive on day one. Google typically takes a cautious approach with consumer-facing generative AI to ensure safety and simplicity.
Monckton notes a precedent with the original Photo Remix launch. That feature initially debuted with only four available presets. Though it expanded to 13 presets months later, it remains far more limited than the freeform editing capabilities found in the underlying Nano Banana model. Users should expect a similar rollout for Soba, likely starting with a handful of curated styles rather than an open-ended chat interface.
Frequently Asked Questions
What is the “Soba” icon in Google Photos?
Soba is a codename for a new AI video tool. Its icon is a YouTube-style play button with a Gemini sparkle logo, found in the “Create” tab of the Android app.
Can I use Soba right now?
No. According to AssembleDebug, the supporting code necessary to run the feature is not yet present in the public app.
What is Gemini Omni?
Gemini Omni is Google’s multimodal video-generation model. It supports conversational editing and multimodal context, meaning it can modify specific events in a video while leaving the rest of the footage untouched.
Do you think limited presets are better for AI tools, or should Google give users full conversational control? Let us know in the comments below or subscribe to our newsletter for the latest AI updates.