wheevu/naver-lens
Embedded AI chat widget for the Naver Smart Store UI. Uses product review data for comprehensive summary. Built for the NAVER Vietnam AI Hackathon.
NAVER Vietnam AI Hackathon 2025 Project
Naver-Lens is an embedded chat that provides organized summaries of product reviews using NAVER's HyperCLOVA AI model.
This codebase includes the Naver Smart Store UI (rebuilt from scratch) and the backend for everything.
Built for the NAVER Vietnam AI Hackathon in the final round.
💡 Note: This is an archived showcase of the project. The live
NAVER Clouddeployment andHyperCLOVAAI services were only available during the hackathon and are no longer running. This project has been set up with a "showcase mode" that uses a saved AI response so you can still see how the full interface works.
What I Did
I built the backend and handled the deployment for our four-person team, team-triple-threat. Here's what I worked on:
- Backend Development: I built the main server using
Node.jsfrom the ground up. - Cloud Deployment: I set up a
Linuxserver onNaver Cloud Platformand got everything running there. I packaged the app usingDockerand handled the full deployment so our project could be accessed online. - Planning & Presentation: I helped come up with the product idea, wrote down what we needed to build, and designed the slides we used to present our project.
How It Works
Backend (Node.js / Express.js)
The backend stores and manages all the product information, lets users filter and browse products, and connects to the AI service (which was developed/tuned by my teammate). I also added caching to make frequently viewed products load faster.
Review Summaries
The main feature is the AI summary panel. When you ask for a summary, the backend streams the response in pieces as it is generated, so you don't have to wait for the entire result.
Database (MongoDB)
I designed the database to handle complex product details like multiple photos, different product options, and customer reviews.
Cloud Deployment (Naver Cloud Platform)
I packaged the entire app using Docker and put it on a live Naver Cloud Platform server so people could actually use it online.
UI
It was Naver's hackathon, so we decided to use one of their existing platforms as the UI. It includes light and dark modes, as well as multilingual support (English/Korean) (the deployed version had working images...).
🇺🇸/🇰🇷 Homepage
Product Catalog & Details
How to Run (Showcase Mode)
1. Clone the repository
git clone https://github.com/wheevu/naver-lens.git
cd naver-lens2. Run the Backend
cd backend
npm install
# Make sure SHOWCASE_MODE=true in .env
npm run dev3. Run the Frontend
cd frontend
npm install
npm run devTech
- Frontend:
React 19,Vite,TypeScript,TailwindCSS,i18next - Backend:
Node.js,Express.js,MVC,Server-Sent Events (SSE) - Database:
MongoDB,Mongoose - AI:
NAVER HyperCLOVA(original),Mock Service(showcase) - Deployment:
Docker,Naver Cloud Platform,Linux (Ubuntu)
I learned git and properly used GitHub for the first time during this hackathon. It was my first time on Slack too.
As you can imagine, it was mostly the same for everything else 😅 (backend, deployment, writing PRs, resolving conflicts, learning how to make a database schema (!),...), over the course of 3 weeks, from ideation to a working deployment.
It was wild. But it was fun.











