A pollen and allergy monitoring platform for Croatia. Real-time pollen data, interactive maps, allergy forecasts, and health recommendations.
Personalized Pollen Alerts
Built a personalized allergen tracking system for alergija.hr, Croatia's pollen monitoring platform.
How it works:
- Users sign in with Google (via Supabase Auth) and select which allergens they're sensitive to — birch, grass, ragweed, and more
- For each allergen, they set a personal threshold level (Low to Extreme) that triggers an alert
- Preferences sync to the cloud, so they persist across devices
- A daily cron job checks real-time pollen levels against each user's thresholds and sends targeted email notifications when levels exceed their personal limits
Tech stack: Next.js, Supabase (Auth + Postgres with Row Level Security), Resend for transactional email, Vercel Cron Jobs
What makes it interesting: The entire feature — 19 user stories, 54 files, 16K+ lines of code including full test coverage — was planned and implemented by an autonomous multi-agent AI pipeline (Antfarm). A planner agent decomposed the task into stories, developer agents implemented each one in fresh sessions, and verifier agents checked the work. Total human intervention: setting up OAuth credentials and running database migrations.
Team
Built together with Tomislav Vulic of Studio Spinalis - sister project to KvalitetaZraka.hr, focused specifically on pollen monitoring and allergy information for the Croatian population.