Skip to content

ai-art-dev99/study-advisor-agent-app

Repository files navigation

StudyAbroad AI — Frontend

The Next.js frontend for StudyAbroad AI: a multi-step agentic advisor that helps international students plan their academic journey to the US, UK, and Canada.

Ask anything — visa requirements, university selection, housing costs, scholarships — and the agent reasons through multiple steps before answering with up-to-date, grounded information.

🔗 Live Demo: [your-deployment-url]
⚙️ Backend (agent + API): study-abroad-assistant repo


What makes this different from a chatbot

Most study abroad tools return static FAQ answers. This agent actually thinks:

User question
    ↓
LangGraph orchestrates multi-step reasoning
    ↓
Claude (Anthropic) decides which tools to call
    ↓
Live web search (Tavily + DuckDuckGo) + RAG over knowledge base (pgvector)
    ↓
Grounded, up-to-date answer streamed back in real time

No hallucinated visa deadlines. No outdated tuition figures. The agent fetches before it answers.


Architecture

┌─────────────────────────────────┐
│   Next.js Frontend (this repo)  │
│   TypeScript · Tailwind CSS     │
│   WebSocket streaming           │
└────────────────┬────────────────┘
                 │ WebSocket
┌────────────────▼────────────────┐
│   FastAPI Backend               │
│   LangGraph · Claude Haiku      │
│   pgvector (Supabase)           │
│   Tavily + DuckDuckGo search    │
│   Deployed on Fly.io            │
└─────────────────────────────────┘

Tech Stack

Layer Technology
Framework Next.js · TypeScript
Styling Tailwind CSS
Real-time WebSocket streaming
Containerisation Docker
Deployment Vercel
Backend (separate) FastAPI · LangGraph · Claude · pgvector

Getting Started

# Clone the repo
git clone https://github.com/ai-art-dev99/study-advisor-agent-app.git
cd study-advisor-agent-app

# Install dependencies
npm install

# Set environment variables
cp .env.example .env
# Add your backend WebSocket URL

# Run locally
npm run dev

Or with Docker:

docker build -t study-advisor-app .
docker run -p 3000:3000 study-advisor-app

Features

  • Real-time streaming — responses stream token by token via WebSocket
  • Multi-domain guidance — universities, visas, housing, scholarships, cost of living
  • Live data — agent fetches current information before answering
  • Clean chat UI — built with Next.js App Router and Tailwind CSS

Related

  • ⚙️ Backend repo — FastAPI, LangGraph, Claude, RAG pipeline

Author

Amirparsa Rouhi · aprouhi.com · LinkedIn

About

Next.js + TypeScript frontend for StudyAbroad AI — a multi-step agentic advisor helping international students plan their academic journey to the US, UK, and Canada.

Topics

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages