Verdict

A mood-based movie discovery application that translates your current state into the perfect watch.

Mobile

Desktop — Curated & Taste Profile

✨ NEW — Movies Search Bar & Search Results Page

The Problem

People often spend more time deciding what to watch than actually watching it. Traditional genre-based filters fail to account for the most important viewing factor: the viewer's current mood and context. Endless scrolling through massive streaming catalogs leads to decision fatigue rather than discovery.

How it works

Verdict introduces the "Tonight" flow:

1. Input: The user selects their current mood, the viewing group size, and available time.
2. Translation: The AI engine interprets these qualitative feelings into precise TMDB Discover API parameters.
3. Curation: Instead of an endless list, the application presents a curated trio from a real candidate pool:
   • Safe: A reliable, highly-rated choice.
   • Risk: The primary, more adventurous recommendation based on the current mood.
   • Wildcard: An unexpected but fitting alternative.

Features

• AI Mood Translation — Converts abstract feelings into concrete search parameters using Groq-powered AI.
• Tri-Choice Curation — Eliminates decision fatigue by limiting options to three distinct categories.
• Taste Profiling — Tracks user interactions (saves, watches, ratings) to build a tailored recommendation profile.
• Localization — Fully internationalized with English and Russian support (via next-intl).
• Daily Lists — Auto-generated daily recommendations based on trending data.

Tech Stack

Backend

• NestJS — Core backend framework providing robust module architecture and scalability.
• Prisma & PostgreSQL — Type-safe database access and schema management.
• Better Auth — Secure and flexible authentication management.
• Groq SDK — High-speed LLM integration for rapid mood analysis and translation.

Frontend

• Next.js (App Router) — React framework optimized for Server-Side Rendering (SSR) and seamless routing.
• React Query & Orval — Type-safe API client generation and robust state management.
• Tailwind CSS & shadcn/ui — Utility-first styling with a tailored dark, amber-themed component library.
• next-intl — Server-side and client-side internationalization.

Infrastructure

• Turborepo — High-performance build system orchestrating the monorepo structure.
• TMDB API — The primary data source for movie metadata.
• Biome — Unified and fast formatting/linting toolchain.

Architecture

The repository utilizes a Turborepo monorepo setup managed via npm workspaces:

• apps/api: The NestJS backend service handling business logic, database operations, and external API integrations (TMDB, Groq).
• apps/web: The Next.js frontend application consuming the API and rendering the user interface.
• packages/: Shared configurations (Biome, TypeScript), utilities, and reusable UI components.

Getting Started

Prerequisites

• Node.js (v20+ recommended)
• npm (v11+)
• PostgreSQL database

Installation

1. Clone the repository:

   git clone https://github.com/phexuss/verdict.git
   cd verdict

2. Install dependencies:

   npm install

3. Configure environment variables in apps/api/.env (and apps/web/.env.local as needed):

   # Example backend configuration
   PORT=3001
   WEB_ORIGIN='http://localhost:3000'
   DATABASE_URL='postgresql://...'
   TMDB_BEARER_TOKEN='your_tmdb_token'
   BETTER_AUTH_SECRET='your_auth_secret'
   BETTER_AUTH_URL='http://localhost:3001'
   GROQ_API_KEY='your_groq_api_key'

4. Setup the database:

   cd apps/api
   npx prisma generate
   npx prisma db push
   cd ../..

5. Start the development environment:

   # Starts both frontend (port 3000) and backend (port 3001) concurrently
   npm run dev

Roadmap

• Advanced Taste Profiling: Improving user-specific recommendations based on historical interaction data.
• Enhanced Group Discovery: Fine-tuning recommendations for multi-person viewing scenarios.
• Streaming Provider Integration: Incorporating platform availability data to filter options by the user's active subscriptions.