Cue
Real-time answers & visuals for video meetings.
About the project
Cue is a silent AI producer for video calls: it joins your Google Meet, listens, and puts what the room needs on its shared screen without ever saying a word. Ask “Cue, pull up last year’s revenue,” or just wonder out loud how big a market is, and within seconds Cue shows a sourced answer, a chart, a photo, a diagram or a clip, like the producer who pulls things up mid-podcast. Under the hood it’s a two-speed brain: Jev, a System 1 decision model on OpenRouter, makes the instant “should I act right now?” call on every pause in about 0.3 s for a fraction of a cent, and only then does Claude (or Gemini, behind one switch) do the slow thinking of researching and drawing the visual. Everything it shows is built to be safe in front of a room: drawn SVGs are sanitized and fully sandboxed, photos are freely licensed and filtered, and web pages are fetched behind address checks so they can’t steer it. You send Cue in from a local control panel that shows what it hears, what it does and what it costs, and it tells the call it’s there with a pinned chat notice.
Built with
AI tools
Prize Categories Entered
- Submitted

Best Use of Gemini API
- Submitted
Best Demo Video Using HyperFrames
We used HyperFrames to make our video in less than 1 hour starting with just the brandkit, brand name, and a readme file describing what is our project. After finishing the first version of the video, we clicked around the video and checked some details of the video then used Framey to fix and make those details smoother and more realistic. Our favorite part about this product is that we can actually point the cursor and pin the part that we want to change in the video and having Framey working concurrently with Claude in making the video was really good
- Event prizeSubmitted
Best Overall Hack
- Event prizeSubmitted
Best Use of System One Models