SentixAi: AI-Powered Movie Review & Sentiment Analysis Platform

SentixAi is a modern, full-stack web application designed for film enthusiasts, analysts, and studios to track real-time audience sentiments. It ingests massive real-world movie review datasets and analyzes them using an embedded, local Artificial Intelligence NLP model.

---

📸 Platform Previews & Screenshots

🎬 Homepage & Movie Discovery

📊 Executive Analytics Dashboard

📽️ Movie Detail & Aspect Sentiment Analysis

🔒 Authentication & Security

Login Screen	Signup & OTP Verification

---

✨ Key Features

• Local AI Sentiment Analysis: Uses DistilBERT (distilbert-base-uncased-finetuned-sst-2-english) via Hugging Face's Transformers.js. Runs entirely on your local machine (zero API costs, complete privacy).
• Real IMDB Data Ingestion: Built-in pipeline to process and ingest a 50,000+ review IMDB dataset directly into the database.
• Social Movie Discovery: A Letterboxd-inspired UI where users can browse popular movies, read reviews, and see aggregate sentiment scores.
• Executive Analytics Dashboard: Visualizes sentiment distribution, total reviews processed, and AI-generated insights.
• Ultra-Secure Authentication: Combines Firebase Auth (Google & Email/Password) with a custom backend-driven 6-digit OTP email verification step (via nodemailer).
• Premium UI/UX: Designed with a dark cinematic theme using React, Tailwind CSS, and a centralized theming system.

---

🏗️ Architecture & Tech Stack

Frontend (/frontend)

• Framework: React 18 + Vite + TypeScript
• Styling: Tailwind CSS + Lucide Icons
• Auth State: Firebase Authentication SDK + React Context
• API Client: Custom fetchWithAuth wrapper for secure backend communication

Backend (/backend)

• Framework: Node.js + Express + TypeScript
• Database: SQLite (local dev.db)
• ORM: Prisma (Type-safe database queries and migrations)
• AI/ML Engine: @xenova/transformers (DistilBERT loaded as a Singleton service)
• Email/Auth: nodemailer for OTP, Firebase Admin SDK for JWT verification

---

🗄️ Database Schema

The relational database is managed via Prisma and includes 8 core models:

• User: Managed alongside Firebase UIDs.
• Movie: Stores metadata and TMDB/IMDb identifiers.
• Review: Raw text and user ratings.
• SentimentAnalysis: 1-to-1 mapping with Reviews, stores AI confidence and Pos/Neg labels.
• OtpVerification: Temporary table for signup email verification.
• IngestionJob: Tracks background dataset processing progress.

---

🚀 Getting Started

Prerequisites

• Node.js (v18 or higher)
• npm or yarn

1. Clone the repository

git clone https://github.com/manishworkss/SentixAi.git
cd SentixAi

2. Setup the Backend

cd backend
npm install

# Setup Prisma and SQLite Database
npx prisma generate
npx prisma db push

# Start the development server
npm run dev

3. Setup the Frontend

Open a new terminal window:

cd frontend
npm install

# Start the Vite development server
npm run dev

4. Access the App

• Frontend: http://localhost:5173
• Backend API: http://localhost:3001
• Prisma Studio (DB Viewer): npx prisma studio (runs on 5555)

---

🧠 How the AI Pipeline Works

1. Ingestion: The admin triggers an ingestion job. The backend reads the IMDB Dataset.csv, cleans the HTML tags, maps them to movies, and saves raw reviews in the database.
2. Background Processing: A non-blocking Node.js service continuously polls the database for unanalyzed reviews.
3. Inference: Reviews are fed into the DistilBERT model in batches of 50.
4. Scoring: The model returns POSITIVE/NEGATIVE labels and a confidence score (0 to 1).
5. Storage & Dashboard: Results are stored in the SentimentAnalysis table and immediately reflected on the frontend Analytics Dashboard.