The Gemini app just got a hugely useful upgrade -- the ability to tell you more about videos you've captured.
Earlier this year, Google introduced a capability for Gemini to analyze YouTube videos. The feature lets you summarize extended content or get a step-by-step overview of a project.
Also: The best AI image generators are getting scary good at things they used to be terrible at
Gemini's video capability is expanding this week; now the AI can take a look at your own camera roll. Just upload a clip to Gemini, and you can get a summary of what's going on and ask extended questions to learn more. You can even ask about what's happening at a certain timestamp or in specific sections of the video.
The feature appears to be available for both Android and iOS, and across multiple models like 2.5 Flash and 2.5 Pro.
ChatGPT and other competitors offer similar capabilities, so it's nice to see Gemini catch up.
Here's how to tell if you have video upload capability for Gemini.
Open your app and tap on the plus sign where you enter your prompt. You can choose to add a video from your gallery or from files. If you can upload videos, you'll be able to do so here. If you can't, videos are grayed out.
When I checked for myself on Gemini 2.5 Flash, I was able to upload videos.
I tried a simple prompt of "tell me about this video" on a 30-second clip I shot of some wild horses I saw on vacation. After about a minute to upload and analyze, Gemini identified where I captured the video (Kill Devil Hills, N.C.), and told me about the overall scene. It went on to explain things like "The video starts with a close-up of a young horse, then pans out to show several other horses of various sizes and colors. They move around the sandy landscape, some walking and some trotting."
I asked Gemini what kind of horses were in the video (I already knew), and Gemini correctly identified the breed and told me a few paragraphs about them. I proceeded to ask a few more questions about my video, and Gemini answered every one in detail.
The feature isn't available to everyone yet, but it is rolling out widely. If you don't have it now, you should soon. Gemini doesn't have the ability to capture video for analysis yet, but that's likely on the way.
Want more stories about AI?Sign up for Innovation, our weekly newsletter.