TL;DR
- We’ve spotted additional details about Gemini’s upcoming video generation capabilities.
- It looks like the feature’s usage may be capped for some users.
- We can now confirm the feature’s codename, providing further evidence of its development.
Last month, we reported that Google might be developing a video generation feature for Gemini within the Google app. While the AI-powered assistant received several upgrades last week, video generation was not among them. However, new evidence suggests the feature is indeed in development and could launch soon.
An APK teardown helps predict features that may arrive on a service in the future based on work-in-progress code. However, it is possible that such predicted features may not make it to a public release.
We’ve discovered references to Gemini’s video generation feature in the latest beta version of the Google app (16.9.39.sa.arm64 beta). Code strings within the app confirms the feature is codenamed “Toucan.”
Notably, the code suggests a potential daily limit on video generation. One line states: “You’ve reached your Toucan generation limit until tomorrow.” It’s unclear whether this restriction will apply to all users or only those on the free version of Gemini, as opposed to Gemini Advanced.
Additionally, we suspect users may be able to choose a dedicated AI model for video generation. Our best guess is Veo 2, which Google introduced last year, promising integration with YouTube Shorts and other products.
Code
Something went wrong. Try again.
Toucan generation was cancelled
Something went wrong. Please try again
Creating your Toucan...
This can take 1-2 mins
You've reached your Toucan generation limit until tomorrow
Get high-quality ones from text with Toucan,
Gemini's new experimental model for generation. Just describe your idea and Toucan can generate
a new one each time you ask.
Icon for the Toucan card
No thanks
Generate stunning ones with Toucan
Try now The code also reveals that Gemini’s experimental video generation feature will rely on text inputs from users. Simply describing an idea will be enough for the AI model to create a video. Additionally, it appears that video generation will take a few minutes to process user prompts.