StoneAgeOS: AI Survival Route Planner

Home Page
Survival Planner Form
Tactical Overview page
Full details Page
History of resutls

Inspiration

I wanted to build a planner that feels like a game UI but solves a real problem: making fast, practical decisions when time, money, and battery are limited. The "futuristic caveman" theme made the experience memorable while still useful.

What it does

StoneAgeOS generates a tactical plan based on:

location (auto-detect, place name, or coordinates)
budget
battery level
available time
mission priorities (food, shelter, safety, charging, comms, workspace, route, tools)

It returns:

best top choice + alternatives
travel distance and ETA
estimated spend and risk level
weather-aware recommendations
fallback plan if the first option fails

Users can also search and reopen past plans from History.

How I built it

Frontend: React + TypeScript + Vite + Tailwind CSS
Backend: Node.js + Express + TypeScript
Data/APIs: Google Places API + OpenStreetMap Overpass fallback, OpenWeather + Open-Meteo fallback, Nominatim/Open-Meteo geocoding
AI: Gemini (with structured JSON schema output)
Database: MongoDB for plan history persistence
Deployment: Frontend on Netlify with separate backend service

The backend pipeline is:

Normalize/validate input
Geocode location
Fetch nearby places + weather in parallel
Score and rank places based on constraints
Generate narrative + structured response with Gemini
Save to MongoDB (graceful fallback if DB is unavailable)

Challenges I ran into

Keeping the response reliable even when third-party APIs fail
Merging different place data sources and deduplicating results
Designing a scoring model that balances distance, budget, safety, weather, and user priorities
Maintaining compatibility between old and new request payload formats
Making the frontend UX rich but still clear and responsive

What I learned

I learned how to design a resilient AI-assisted system with strong fallbacks, schema-constrained LLM output, and practical production concerns like validation, persistence failure handling, and API timeout management.