A short YouTube demo by Emory Fierlinger exploring a Generative UI prototype that combines OpenAI's real-time voice API with Google’s Nano Banana image model to assist a user planning a cherry-blossom-season trip to Japan. The demo shows conversational planning (cities, timing, a sample 5-day Tokyo–Osaka itinerary), image generation of Tokyo sakura scenes, and a brief look at flight-search guidance and cost-estimation heuristics. The narrator frames this as an improved UX for voice agents compared with current simplistic visualizers.
Chapters
Recently, I've been exploring Generative UI for user interfaces. This is an example that uses the OpenAI real time voice API and the new Google Nano Banana model. Hey, how's it going? I'm doing great, thanks for asking. How about you? Yeah, well, thanks. Hey, I was wondering. I want to plan a trip to Japan and I want to go during cherry blossom season. Can you help me plan the trip? Sure, I'd love to help with that. Cherry blossom season in Japan is beautiful, but the timing can vary depending on the region. What cities are you thinking of visiting? I don't know. Maybe you can help me explore that. Of course. Let me gather some information about the best spots and times for cherry blossoms in Japan. Just a moment. Some of the top cherry blossom spots in Japan are Tokyo and Kyoto in early April, Osaka around mid April, and Sapporo in late April. Each of these places. Hey, do you think you could show me some photos of Tokyo during cherry blossom season? Let me find some beautiful examples for you. One moment. Here's a beautiful view of Tokyo during cherry blossom season. The pink sakura flowers look amazing with the city backdrop. That's awesome. Hey, do you think you could plan a five day trip for me? And I want to go to Osaka as well. Let me put together a sample 5 day itinerary for Tokyo and Osaka during cherry blossom season. One moment. Here's a sample 5 day itinerary. Days 1 to 3 in Tokyo, visit Ueno Park, Shinjuku Gyoen and Chidori Gafuji for for amazing cherry blossom views. Days four to five in Osaka. That's great. Hey, do you think you could also help me find some return flights from London for that around that time? Let me check the flight options from London to Japan around cherry blossom season. Just a moment. Essentially, what's happening in the background is it's using the OpenAI responses API to do a web search tool call and then it's combining the data from the web search tool call to call Google's Nano Banana image generation model and come back with a generated ui. To find return flights from London to Japan during cherry blossom season, it's best to book early, be flexible with your dates, and consider flying into either Tokyo or. Thanks. How much do you think it would roughly cost me for a return flight? Let me see if I can get a rough estimate on return flight costs for that time. Time. Just a moment. So as you can see, it's essentially just a better way to interact with voice agents. Because right now all you can do. All you can do at the moment is just look at a circle that's dancing around your screen. And I think this is a way better example of what the future of Generative UI could look like. Thanks.
