City Hunter

City Hunter turns sightseeing into a real-world RPG. Earn XP by exploring landmarks and completing AI-driven quests. Uncover hidden history and challenge friends. Don't just visit Play the City.

Comment

CityHunter: The Smart Compass for the Solo Explorer

"Don’t just visit. Play the City."

CityHunter is a gamified urban exploration ecosystem that solves the two biggest problems of solo travel: loneliness and paralysis of choice. By replacing static guidebooks with a conversational, game-based interface, CityHunter turns a simple walk into a meaningful quest.

🌍 The Problem & The Vision

Have you ever visited a landmark and forgot everything about it a week later? Or worse—do you live in a city full of secrets you’ve never even noticed?

Traditional travel is passive. We follow a blue dot, read a dry plaque, and move on. We look, but we don’t see. CityHunter was built on the philosophy that learning is a byproduct of play. It bridges the gap between "visiting" a city and actually "solving" it.

🚀 The Core Experience

1. Two Roles, One City

The Hunter (User): The front-line explorer. Users wander with purpose, solving location-based mysteries, gaining XP, and unlocking real-world rewards like local shop discounts.

The Architect (Admin): The curator. Admins have the power to manage "Secrets," design optimized "Walks" (parcours), and deploy live community events to keep the city evolving.

2. Gamified Utility

Quests & XP: Every monument is a mission. Decipher riddles and verify your location to level up.

The Walk Engine: Curated routes that guide you step-by-step, ensuring you maximize sights and never feel lost.

Social Connectivity: Join "Squads," see active hunters on the map, and compete on global leaderboards.

🧠 Agentic Intelligence: Powered by Gemini

At the heart of CityHunter is Monumentum, a sophisticated multi-agent system built on Google's Agentic Development Kit (ADK) and powered by Gemini 3 Flash.

Why Gemini 3 Flash?

A travel companion must respond in real-time—latency breaks immersion. We chose Gemini 3 Flash for its incredible speed and reasoning capabilities, allowing the AI to "think" and query geospatial databases instantly while maintaining a witty, conversational flow.

Hierarchical Agent Architecture

The system uses a Hierarchical Agent Pattern where the root agent delegates tasks to specialists:

🤖 The Root Agent (Monumentum): The "Concierge" that handles greetings, intent analysis, and routing.

🏛️ Sub-Agent A (MonumentExpert): The "Historian." It uses custom tools like search_monuments(lat, lon) to perform geospatial queries on our MongoDB database, challenging users to notice architectural details.

🗺️ Sub-Agent B (TravelManager): The "Logistics Guide." It focuses on route planning, vibes, and duration using the search_walks tool.

Key Capabilities

Context-Aware RAG: Monumentum is aware of your exact coordinates. When you ask "What's that building?", the agent performs a real-time geospatial lookup before answering.

Strict Tool Usage: Prioritizes verified "CityHunter Certified" data over general web searches to ensure high-quality, authentic insights.

🛠️ Tech Stack

Frontend: Next.js (React), Tailwind CSS, Framer Motion for immersive animations.

Backend: FastAPI (Python) for a high-performance, asynchronous API.

AI/Agents: Google GenAI SDK & Google ADK (Agentic Development Kit).

Database: MongoDB Atlas with GeoJSON support for complex spatial queries.

Infrastructure: Dockerized deployments on Google Cloud Run.

📽️ Demo Narrative

The Hunter Journey: Login as a Hunter and let the radar pulse. Approach a monument, trigger the "Active Sightseeing" AI, and solve the mystery. Watch your XP ring fill and your rewards unlock.

The Architect Control: Login as an Architect to see the city from above. Drag-and-drop pins to create a new "Revolutionary Trail" and watch in real-time as Hunters begin the quest you just built.

"Wander with Purpose."

Built With

Share this project:

Updates