Launching Seoul 3D Atlas, an interactive 3D experience for exploring how Seoul looks, moves, and changes. Built iteratively with GPT-6 Astra, it combines city-scale 3D visualization, transit, tourism, multilingual UX, and environmental effects in one web experience.
What became possible in your product with Astra that was not practical before?
Maker
Astra made it practical for me to build and continuously evolve a city-scale product as a solo developer.
Before Astra, combining hundreds of thousands of buildings, transportation, tourism data, multilingual UX, weather, seasons, and interactive 3D features would have required much more time or a larger team.
The biggest change was iteration speed. Astra helped me understand the existing codebase, implement features, debug regressions, connect data sources, and continuously refine the UX.
This speed allowed Seoul 3D Atlas to grow from a simple visualization into a much broader interactive city experience. When I shared the project on X, it received 330K+ views—an achievement that demonstrated strong public interest in this new way of exploring a city.
Report
Maker
📌
Hi Product Hunt! 👋
I’m Jinho, the maker of Seoul 3D Atlas. This project started with a simple question: what if Seoul could be explored as a living 3D city instead of a flat map?
The first version focused on visualizing the city itself. From there, the project kept evolving through repeated use and feedback. I added subway and bus routes, landmarks, hotels, cultural sites, multilingual navigation, multiple currencies, weather, seasons, and day/night experiences.
Today, Seoul 3D Atlas brings together more than 267,000 buildings and multiple layers of public city data in one interactive web experience. One of the most unexpected moments came when I shared the project on X (Twitter). The post and demo reached more than 330,000 views, and the reactions helped convince me that people were genuinely interested in exploring a city this way.
What makes this launch especially meaningful to me is how it was built. I used the latest version of GPT-6 Astra throughout the development process—not just to generate an initial prototype, but as part of an ongoing engineering workflow. I repeatedly defined new requirements, tested results, found problems, refined the UX, debugged implementation issues, and evolved the product over many iterations.
For me, the most interesting part of working with Astra was seeing how AI-assisted development changes when the project keeps growing and the requirements keep changing.
Seoul 3D Atlas is still an experiment rather than a perfect digital twin. My goal is to explore a new way of experiencing a city by combining 3D visualization, transportation, tourism, and urban data in one place.
I’d love to hear what you think—and what you’d explore first: transit, landmarks, tourism, weather, or just flying around Seoul. 🚇🏙️🌧️
Thanks for checking it out!
Report
Maker
Hi I’m excited to share that JoongAng Ilbo’s Factpl Magazine in South Korea has featured my project, Seoul in 3D Miniature, as a real-world application of GPT-6 Astra.
This interactive 3D map allows users to explore all 25 districts of Seoul and approximately 267,000 buildings directly in a web browser. The initial version was built with Astra in about 43 minutes. Since then, I have continued expanding the project as new requirements emerged through real-world use.
It began as a 3D visualization of Seoul. Over time, I added automated landmark fly-throughs, subway and bus route exploration, tourist attraction and hotel information, apartment sales and rental data, multilingual support, and much more.
The most fascinating part of this project was seeing how the role of generative AI is changing. AI is no longer limited to answering questions or generating small pieces of code. A new approach to software development is emerging—one in which AI can understand a user’s goal, break it down into tasks, connect data sources, write code, and build a working service.
This project did not begin with a fully completed design. It evolved into its current form through a repeated cycle of using the product, discovering the next requirement, implementing it, and testing it again.
I believe this way of working will become increasingly important in AI-powered software development.
The focus is shifting from “What should we code?” to “What problem should we define, and how quickly can we validate and improve the solution together with AI?”
Being featured by JoongAng Ilbo was personally meaningful to me. More importantly, however, I see this project as an example of how rapidly AI is transforming the way ideas can be turned into real products.
Hi I’m excited to share that JoongAng Ilbo’s Factpl Magazine in South Korea has featured my project, Seoul in 3D Miniature, as a real-world application of GPT-6 Astra.
This interactive 3D map allows users to explore all 25 districts of Seoul and approximately 267,000 buildings directly in a web browser. The initial version was built with Astra in about 43 minutes. Since then, I have continued expanding the project as new requirements emerged through real-world use.
It began as a 3D visualization of Seoul. Over time, I added automated landmark fly-throughs, subway and bus route exploration, tourist attraction and hotel information, apartment sales and rental data, multilingual support, and much more.
The most fascinating part of this project was seeing how the role of generative AI is changing. AI is no longer limited to answering questions or generating small pieces of code. A new approach to software development is emerging—one in which AI can understand a user’s goal, break it down into tasks, connect data sources, write code, and build a working service.
This project did not begin with a fully completed design. It evolved into its current form through a repeated cycle of using the product, discovering the next requirement, implementing it, and testing it again.
I believe this way of working will become increasingly important in AI-powered software development.
The focus is shifting from “What should we code?” to “What problem should we define, and how quickly can we validate and improve the solution together with AI?”
Being featured by JoongAng Ilbo was personally meaningful to me. More importantly, however, I see this project as an example of how rapidly AI is transforming the way ideas can be turned into real products.
reference: https://www.joongang.co.kr/article/25462714