AI-Assisted FAQ & Query Resolution System

An intelligent knowledge management platform built for internship programs. Interns get instant AI-powered answers to their questions; admins manage the knowledge base, moderate discussions, and keep the platform growing.

---

What It Does

• Interns ask questions → AI searches the FAQ knowledge base and returns an instant answer
• If no answer exists, the query is saved and posted to the Discussion Room for the community to answer
• Admins moderate answers, publish official FAQs, and manage users
• Every published FAQ makes the system smarter for future users

---

Features

For Interns

• FAQ Hub — Browse all published FAQs across 14 categories
• Hybrid Search — Semantic + keyword search that understands intent, not just keywords
• AI Chatbot — Floating assistant that answers questions using the FAQ knowledge base
• Discussion Room — Ask questions, post answers, upvote/downvote community responses
• Points & Leaderboard — Earn points for contributing; compete with other interns
• Real-time Notifications — Get notified when someone answers your question or upvotes you
• Programme Overview — Live-scraped overview of the internship programme

For Admins

• Moderation Panel — Accept answers, delete posts, ban users from Discussion Room
• FAQ Publishing — Manually create and publish FAQs to the knowledge base
• User Management — View all users, promote interns to admin, manage roles
• Analytics Dashboard — Total users, questions, answers, FAQs at a glance

For Super Admins

• Everything admins can do
• FAQ Manager — Full control over the FAQ database
• Audit Log — Track all admin actions on the platform
• Assign and revoke admin roles

---

Tech Stack

Layer	Technology
Frontend	React.js, CSS Variables, custom design system
Backend	Node.js, Express.js
Database	MongoDB Atlas
AI — Embeddings	Google Gemini
AI — Responses	Groq LLM
Real-time	Socket.IO
Auth	JWT (access + refresh tokens), Google OAuth
Search	Hybrid semantic + keyword with in-memory embedding cache

---

Getting Started

Prerequisites

• Node.js v18+
• MongoDB Atlas account (or local MongoDB)
• Gemini API key
• Groq API key
• Google OAuth credentials

Installation

1. Clone the repo

git clone <repo-url>
cd project-root

2. Install dependencies

# Backend
cd server
npm install

# Frontend
cd ../client
npm install

3. Configure environment variables

Create server/.env:

PORT=3000
NODE_ENV=development
MONGODB_URI=your_mongodb_atlas_connection_string
JWT_ACCESS_SECRET=your_access_secret
JWT_REFRESH_SECRET=your_refresh_secret
GOOGLE_CLIENT_ID=your_google_client_id
GOOGLE_CLIENT_SECRET=your_google_client_secret
CLIENT_URL=http://localhost:5173
CLIENT_URLS=http://localhost:5173,http://localhost:5174
LLM_ENDPOINT=https://api.openai.com/v1/chat/completions
LLM_API_KEY=your_llm_api_key
LLM_MODEL=gpt-4o-mini
CHROMA_HOST=localhost
CHROMA_PORT=8000
CHROMA_COLLECTION=faq_embeddings
POINTS_JOB_INTERVAL_MS=3600000

Create client/.env:

VITE_API_BASE_URL=http://localhost:3000/api
VITE_SOCKET_URL=http://localhost:3000

4. Run the project

Open two terminals:

# Terminal 1 — Backend
cd server
node server.js

# Terminal 2 — Frontend
cd client
npm run dev

Visit http://localhost:5173

---

Roles

Role	Access
intern	FAQ Hub, Search, Chatbot, Discussion Room, Ask Question, My Questions, Leaderboard
admin	All intern views + Admin Area (moderation, FAQ publishing, user management, analytics)
super_admin	All admin capabilities + FAQ Manager, Audit Log, full role management

To make a user an admin, update their role field in MongoDB from intern to admin.

---

Project Structure

project-root/
├── client/               # React frontend
│   └── src/
│       ├── api/          # Axios instance + search API
│       ├── components/   # Reusable components + layout
│       ├── context/      # Auth context
│       ├── pages/        # All page components
│       ├── services/     # API service functions
│       └── styles/       # CSS design system
│
└── server/               # Express backend
    ├── config/           # Database connection
    ├── controllers/      # Route handlers
    ├── middleware/       # Auth + error handling
    ├── models/           # Mongoose schemas
    ├── routes/           # API routes
    └── services/         # Search, embeddings, scraper

---

API Overview

Method	Endpoint	Description
POST	/api/auth/login	Login with email/password
POST	/api/auth/register	Register new account
POST	/api/auth/google	Google OAuth sign-in
GET	/api/faqs	Get all published FAQs
GET	/api/search	Hybrid search across FAQs
POST	/api/queries	Save unresolved query
GET	/api/questions	Get all Discussion Room questions
POST	/api/questions	Post a new question
GET	/api/questions/:id	Get question with answers
POST	/api/answers	Post an answer
POST	/api/answers/:id/vote	Upvote or downvote an answer
GET	/api/admin/stats	Get platform analytics
GET	/api/admin/users	Get all users
PUT	/api/admin/users/:id/role	Update user role
GET	/api/internship/overview	Get programme overview

---

Vercel Deployment (Client)

This repository is a split architecture:

• client is a Vite SPA suitable for Vercel static hosting.
• server is a long-running Node + Socket.IO API and should be hosted on a persistent backend platform (Render, Railway, Fly, VM, etc.), not Vercel serverless.

Vercel project settings

• Framework Preset: Vite
• Root Directory: client
• Install Command: npm install
• Build Command: npm run build
• Output Directory: dist

Required Vercel environment variables (client)

• VITE_API_BASE_URL=https://<your-backend-domain>/api
• VITE_SOCKET_URL=https://<your-backend-domain>

The SPA fallback rewrite is configured in client/vercel.json and excludes /api and /socket.io paths so API traffic is not rewritten to HTML.

---

Points System

Action	Points
Post an answer	+10
Receive an upvote	+5
Receive a downvote	-2
Answer accepted by admin	+20

---

What's Not Built (Future)

• Document upload via admin dashboard (PDF/TXT/DOCX → auto FAQ extraction)
• RAG pipeline for uploaded documents
• Query clustering
• n8n automation workflows
• Voice-based queries
• Mobile app