UVA Chatbot

A chatbot designed for the UVA Career Center

uva_chatbot_compressed.mp4

How It Works

The site is built with Next.js and consists of a single page. When a user sends a message, it triggers a POST request to api/chat. This request is handled by route.js, which connects to Gemini, streams the response, and sends it back to the user in real time.

src
├── app
│   ├── api
│   │   └── chat
│   │       └── route.js
│   ├── globals.css
│   ├── layout.jsx
│   └── page.jsx
├── components
│   └── Loader.jsx
└── lib
    └── utils.js

To-Do List

Currently, the model sends a single request without maintaining context. The following improvements are planned:

• Integrate RAG (Retrieval-Augmented Generation): Use a document store like Pinecone to provide contextual responses.
• Add Pricing Estimates: Build a pricing table to calculate and display usage-based costs.
• Select Hosting Provider: Research and choose a reliable and scalable hosting solution.
• Domain Integration: Connect the app to the UVA Career Center domain for deployment.