← Selected work

Case study

Atmos Weather

Personal project · 2026 · Built with AI using Codex

Next.js · React · TypeScript · Open-Meteo · MapLibre · Recharts

Built with AI using Codex. Atmos is a personal project exploring AI-assisted web development.

Overview

Atmos brings everyday forecasts and deeper weather exploration into one responsive dashboard. Powered by real Open-Meteo data, it combines current conditions, hourly detail and a 16-day outlook with interactive maps, observed radar and specialist tools.

Atmos weather dashboard showing current conditions, hourly forecasts and charts for Bruges

Key features

Everyday forecasts

Current conditions, hourly charts and a 16-day forecast, with worldwide location search, saved places and comparisons.

Maps & observed radar

Interactive MapLibre maps with sampled forecast layers and wind animation, plus optional RainViewer playback of observed radar.

Beyond the forecast

Air quality, marine conditions, solar energy, model comparison, historical weather and climate exploration.

Personal & portable

Responsive layouts, light and dark themes, configurable units and an installable PWA with saved forecasts available offline.

Technical stack

Next.js and React provide the application and server API routes, with TypeScript throughout. TanStack Query handles remote data, Zustand stores preferences, and Zod validates API responses. MapLibre powers the maps and Recharts renders the forecast charts.

Explore the code behind Atmos.

View source on GitHub ↗