Movie Night Picker

Live SiteBuilt with AI

Stop arguing about what to watch. Movie night decisions in 4 minutes, not 40.

AI Tools Used

ChatGPT

Ideation, strategy, UI/UX design, code prompt generation

Claude Code

Full implementation from prompts

Workflow

I defined the room flow, tie logic, and "watched" logging. AI generated the base UI + API stubs; I iterated to a shippable demo.

⏱️Timeline: MVP in 3 hours, polished version in 3 days total.

The Problem

My friend group would spend 30-40 minutes debating what movie to watch. Everyone had opinions, nobody wanted to be "the decider," and we'd end up doom-scrolling Netflix until someone gave up. I wanted a fair, fast way to pool suggestions under a theme and let chance pick the winner - no hurt feelings, no endless debate.

What I Built

A group movie picker: create a themed room, friends join and submit movies, hit "Pick Random Movie" for an animated selection, see where to stream it, and log it to your watched history. Firebase keeps everything synced in real-time so everyone sees updates live.

Real Usage

4 friends in my immediate movie-watching crew use it. We've used it for themed nights like Animated Movies. Group size ranges from 2-6 people. One group now uses it every Friday.

Architecture

Tech Stack

Frontend

React

Backend

Node.js

AI/ML

What I Learned

Firebase setup

Not trivial, but critical to get right. Real-time syncing lets people use the app simultaneously without conflicts - essential for the group experience.

Communicating UI to AI

Describing visual design in words and screenshots to ChatGPT/Claude was trickier than expected. Iterations helped, but precision matters.

Animation complexity

Had to break animations into individual frames and phases (scroll → jiggle → spotlight → modal). Building too much at once caused breaks that neither I nor Claude Code could debug well. Incremental is the way.

Product intuition validated

I'd been thinking about this idea for a while. Turns out all the features I imagined are actually getting used - nothing felt like bloat.

Where It Could Go

Next Features

✓**AI recommendations:** Suggest similar movies based on your theme - "You picked a heist movie, here are 5 more"
✓**Natural language constraints:** "We have 2 hours max and want something laid-back" → AI translates that into smart filtering before the random pick
✓**LLM plot summaries:** Quick context on each movie before you commit

Bigger Vision

This pattern works beyond movies. Groups want easy ways to pool suggestions and fairly pick winners for restaurants, travel destinations, weekend plans - anything where collective decision fatigue exists. Could expand into a general "group decision engine."