Reverix 🎬

AI-Powered Smart Movie Booking Platform

"Tell me your mood — Reverix finds your perfect movie, theatre, and seat."

---

What is Reverix?

Reverix is a smart movie booking platform that goes beyond just listing movies and seats. It understands your mood, your group, and recommends the perfect movie + theatre + seat zone — powered by AI.

Think BookMyShow meets an AI concierge.

---

Screenshots

Home — Cinema, tuned to your mood

AI Picks — Mood input → Ranked results

---

Features

AI Mood-Based Recommendation

Tell Reverix how you feel in plain English:

• "Feeling excited, going with 5 friends"
• "Romantic evening with my partner"
• "Family of 7, want something light"

Reverix calls Llama 3.3 70B via OpenRouter to analyze your mood, match it to currently playing movies, recommend theatres by vibe, and suggest the best seat zone.

Theatre Vibe Classification

Every theatre is classified by its personality:

Vibe	Best For
CELEBRATION	Group of friends, mass entertainers
SILENT	Solo viewers, art films
FAMILY	Families, animation, drama
DATE_NIGHT	Couples, romance
PREMIERE	IMAX, blockbusters

Group-Aware Seat Recommendation

Group Type	Recommended Zone	Reason
Friends (4+)	MIDDLE	Best group experience
Family	BACK	Spacious, easy exit
Couple	MIDDLE	Perfect view
Solo	MIDDLE	Best screen experience

CinePrime Subscription

Premium users get early access to tickets before general public — implemented with FIFO queue logic and role-based access control.

Real Movie Data

Movies synced from TMDb API with automatic mood-tag generation based on genre classification.

---

Tech Stack

Layer	Technology
Language	Kotlin 1.9.22
Framework	Spring Boot 3.2.3
Security	Spring Security + JWT
Database	MySQL 8
Migrations	Liquibase
AI	OpenRouter + Llama 3.3 70B (free)
Movies API	TMDb

---

Architecture

Client Request
     │
     ▼
Spring Security (JWT Filter)
     │
     ▼
REST Controllers
     │
     ├── AuthController           → Register / Login
     ├── MovieController          → TMDb sync + listing
     ├── TheatreController        → Vibe-based recommendation
     ├── RecommendationController → AI mood engine
     └── BookingController        → Seat lock + confirm
     │
     ▼
Service Layer
     │
     ├── RecommendationService → OpenRouter/Llama AI call
     ├── TmdbService           → TMDb API sync
     ├── BookingService        → FIFO queue + seat locking
     └── TheatreService        → Vibe matching algorithm
     │
     ▼
MySQL Database (Liquibase migrations)
     │
     └── users / theatres / movies / shows / seats / bookings

---

API Endpoints

Auth (Public)

POST /api/auth/register    — Create account
POST /api/auth/login       — Get JWT token

Movies (Public)

GET /api/movies/now-playing          — Currently showing
GET /api/movies/popular              — Popular movies
GET /api/movies/search?query=        — Search movies
GET /api/movies/{id}                 — Movie details
GET /api/movies/rentable             — Movies to rent

Theatres (Public)

GET /api/theatres                             — All theatres
GET /api/theatres/city/{city}                — By city
GET /api/theatres/city/{city}/vibe/{vibe}    — By vibe
GET /api/theatres/recommend?city=&groupType= — Smart recommend

AI Recommendation (Public)

POST /api/recommend
Body: {
  "mood": "excited, going with friends",
  "city": "Chennai",
  "groupType": "friends",
  "groupSize": 5,
  "preferredZone": "MIDDLE"
}

Bookings (JWT Required)

POST   /api/bookings/lock-seats       — Lock seats (10 min)
POST   /api/bookings/confirm          — Confirm booking
DELETE /api/bookings/{id}             — Cancel booking
GET    /api/bookings/my-bookings      — Your bookings
GET    /api/bookings/recommend-seats  — Smart seat picker

---

Local Setup

Prerequisites

• Java 17
• MySQL 8
• Gradle

Steps

# Clone the repo
git clone https://github.com/ashish-babu-03/Reverix.git
cd Reverix

# Set these in src/main/resources/application.properties
spring.datasource.url=jdbc:mysql://localhost:3306/reverix_db?createDatabaseIfNotExist=true
spring.datasource.username=root
spring.datasource.password=yourpassword
tmdb.api.key=your_tmdb_key
openrouter.api.key=your_openrouter_key
openrouter.model=meta-llama/llama-3.3-70b-instruct:free
app.jwt.secret=your_secret_key

# Run — Liquibase creates all tables automatically
./gradlew bootRun

Server starts at http://localhost:8080

---

Key Design Decisions

Why Liquibase over Hibernate auto-create? Version-controlled migrations mean any developer can clone and run with zero manual DB setup. Each changeSet is independently trackable and reversible.

Why OpenRouter + Llama instead of GPT-4? Llama 3.3 70B on OpenRouter is completely free for development, follows the same OpenAI-compatible API contract, and can be swapped for any other LLM by changing one config line.

Why ENUM columns for theatre vibe and seat zone? Enforces data integrity at the database level — no invalid vibe types can ever be inserted, even by direct DB access.

Why 10-minute seat lock window? Mirrors real booking platforms — long enough to complete payment, short enough to prevent seat hoarding.

---

What I Learned

• Designing multi-entity REST APIs with proper separation of concerns
• Integrating third-party APIs (TMDb, OpenRouter) with graceful fallback handling
• Implementing time-based resource locking (seat expiry)
• Version-controlled database migrations with Liquibase
• Role-based access control with Spring Security + JWT

---

Author

Ashish Babu Z

• Email: ashish.babu.sde@gmail.com
• LinkedIn: linkedin.com/in/ashish-babu-z
• GitHub: github.com/ashish-babu-03

---

Built as a portfolio project to demonstrate Spring Boot + Kotlin backend engineering with AI integration.