AI-powered outfit recommendations based on hyper-local weather data. Never overdress or underdress again.

Production domain: https://skystyle.app

• Multi-source weather — aggregates data from OpenWeatherMap, Open-Meteo, and Bureau of Meteorology (Australia) for accuracy, with hourly forecasts
• AI outfit recommendations — uses OpenAI GPT-4o or Google Gemini to suggest what to wear
• Follow-ups — ask follow-up questions like "should I bring an umbrella?" or "what if I need formal shoes?"
• Digital closet — add your wardrobe items so the AI knows what you own
• Bring Your Own Key — Pro users can use their own AI API key (never stored)
• Custom weather sources — Pro users can add their own weather data sources
• GPS & manual location — use your browser's location or search for any city
• Dark mode — automatic, based on system preference

Dashboard	Weather Flow

Usage + Limits	Account + Security

This repository is now organized as a single monorepo for multiple deployable projects.

/
├── apps/
│   ├── web/        # Sky Style main app (Next.js)
│   ├── docs/       # VitePress docs website
│   └── api/        # Future standalone API placeholder (api.skystyle.app)
├── packages/       # Shared packages for future reuse
├── supabase/       # Database schema and SQL assets
└── ...root config files

Project targets:

• apps/web -> skystyle.app (or current preview domain)
• apps/docs -> https://docs.skystyle.app
• apps/api -> api.skystyle.app (future)

• Next.js 16 (App Router + Turbopack)
• Supabase (Postgres + Auth adapter)
• NextAuth v5 (GitHub + Google OAuth)
• OpenAI / Google Gemini (AI)
• OpenWeatherMap + Open-Meteo (weather)
• Tailwind CSS 4
• Vercel (hosting + analytics)

See SETUP.md for full deployment and local development instructions.

Quick start:

cp .env.example apps/web/.env.local # or use vercel env pull
# fill in values in apps/web/.env.local
npm install
npm run dev

Docs quick start:

npm run docs:dev

• SSRF protection — custom weather source URLs are validated against private/internal hosts
• HTTPS only — custom source URLs must use HTTPS
• No secrets in client — all API keys are server-side only
• BYOK keys are never stored — user-provided AI API keys are used for the single request and discarded
• Row Level Security — Supabase RLS policies restrict data access to the owning user
• JWT sessions — NextAuth uses JWT strategy; the database adapter is wrapped in a safe fallback
• Input validation — coordinates, API inputs, and user content are validated before use
• Rate limiting — daily usage limits prevent abuse of AI and weather APIs

Issues and pull requests are welcome. Please be respectful and follow the People First Design principles.

• Terms of Service
• Privacy Policy

Both follow the People First Design principles — plain language, minimum data collection, no dark patterns.

See LICENSE.

If you like my project and want to support the development:

☕ Buy me a coffee
https://buymeacoffee.com/coolmanyt

:)