Mohammad El-Musleh

Mohammad El-Musleh

(Embedded) Software Engineer

Go Fish: AI-Powered Group Activity Planner

Published:

🤖 AI Disclosure: The content here is created with AI assistance and reviewed by me. I believe in transparency about how content is made!

Background

Go Fish was built at the Cursor Hackathon Heilbronn (March 2026), organized by CREATORS and Arkadia — a 30-hour sprint alongside 66 teams from across Europe. The event was centered on <a href=”/skills/?tag=Vibe+Coding”
class=”cv-skill-tag”
data-skill=”Vibe Coding”>Vibe Coding</a>
using AI-assisted development (primarily Cursor) as the main methodology, not just a helper.

The idea came from a simple frustration: group chats full of “we should do something” that never turn into actual plans because no one owns the next step.

What I Built

Go Fish is a web app where one person creates an event and shares a single invite link with the group. The flow from there:

  1. Create an event with a title, description, city, and response window
  2. Share one invite link with the group
  3. Collect availability and preference benchmarks from each participant
  4. Generate ranked activity suggestions via an LLM (OpenRouter API)
  5. Finalize by picking an option and sending an email confirmation to all participants

The output is a small shortlist of suggested activities with reasoning, not an open-ended list to argue about.

Tech Stack

  • Backend: Express (Node.js), port 3000
  • Frontend: React (Vite), port 5173
  • Database: PostgreSQL (Docker)
  • AI: OpenRouter API for LLM-generated suggestions
  • Email: Resend API for group confirmation emails
  • Optional enrichment: Google Places, OpenWeatherMap, Foursquare, Ticketmaster (API keys, not core to the flow)
  • Deployment: Railway (backend + frontend), also available at go-fish-nu.vercel.app

Honest Scope

This was built in 30 hours, so there are rough edges. The Ticketmaster integration is wired up as an optional data source, not a direct booking system — the app confirms plans via email, it doesn’t purchase tickets. The LLM-powered suggestions work well for the hackathon demo; a production version would benefit from better preference modeling and real-time venue data.

Explore the full codebase: github.com/el-musleh/Go-Fish
See all hackathon projects: creators-ecosystem.de/de/hackathon-projects