It’s no secret that Google Gemini has been getting smarter since its launch. It can answer questions, summarize documents, generate visuals, and even research the web for you. Now, Gemini is adding a new feature that allows users to upload and analyze videos. This will come in handy for those who want to extract insights or answers from videos, such as lectures or other content. It’s still in development, but AssembleDebug managed to give us a hands-on demo showing how it works.
Someone managed to get Gemini to analyze a video, and this could be your first look
Android Authority somehow managed to activate the functionality in Gemini v16.15.38.sa.arm64 beta that now lets analyze videos. The video analysis feature allows users to upload videos directly into the Gemini app, where the AI will then analyze the footage and answer questions about the content. For example, you may ask Gemini to identify the location shown in a video, summarize a tutorial, or even offer insights on a specific scene.
One such demo showed Gemini successfully identifying the location of a video just by analyzing the architecture in the background, without any context being provided. This demonstrates the potential of AI to understand and interpret video content in a smart and useful way.
However, there are a few limitations to keep in mind. Currently, Gemini can only handle videos that are up to 5 minutes long. If a video exceeds this length, users will receive an error message. Additionally, while you can upload multiple videos at once, the total combined length must not exceed the 5-minute limit.
This feature could be extremely useful in a variety of situations. It could help users get summaries of video tutorials, analyze Zoom call recordings, and even review recorded lectures. Although it’s not clear when this feature will be officially launched, there’s a good chance it will be unveiled during Google I/O 2025, which takes place next month.
