Forkcast: Conquering Group Dining Decisions
🍽️ The Problem
We've all been there—sitting with friends, endlessly debating where to eat. Everyone has unique preferences, dietary restrictions, and budgets. More often than not, we'd give up and order separately, missing out on new culinary adventures.
🤖 The Solution
With 2025 being dubbed "the year of agents," I built Forkcast—a multi-agent system that orchestrates group dining decisions seamlessly.
🛠️ The Journey
After researching existing solutions and identifying gaps, I dove deep into:
- Google AI SDK documentation and multi-agent architectures
- Model Context Protocol for agent communication
- Google Maps API for location intelligence
- Firebase ecosystem for auth, hosting, and data storage
🎯 Key Challenges
The toughest hurdle? Deployment. After weeks of local testing, configuring the production environment felt overwhelming. Thankfully, Google Cloud Run provided the flexibility I needed.
🚀 The Takeaway
Beyond building a "cool ass project," these 30 days of hacking taught me cutting-edge technologies and easily the most fun I've had coding this year !
The learnings and rush of building something meaningful—that's the real victory.
Built With
- angular.js
- cloudrun
- firebase
- firestore
- foursquare
- google-gmail-oauth
- google-maps
- googleadk
- mcp
- openapi
- yelp-fusion


Log in or sign up for Devpost to join the conversation.