Wind-optimised cycling route recommendations so you always have a tailwind on the way home.
Live site - Built for cyclists in and around Leighton Buzzard, UK.
Tailwise takes a library of cycling routes and ranks them based on real-time wind conditions. For each route, it calculates whether you're better off riding it as-planned or in reverse, and tells you how strong the tailwind advantage is on the homeward leg.
It also integrates with Strava, pins cafe stops along the way, and has a dedicated portal for Leighton Buzzard Road Cycling Club weekly rides.
Weather is a huge factor in cycling, particularly the wind. Having a nice tailwind on the way home makes a big difference to the enjoyment of a ride. I've watched and listened to local riders planning routes and it starts with the wind direction and looking at a big list of potential routes, I thought I could make that easier.
Built as a personal project for me and other friends, then extended for my local cycling club.
- Wind-optimised recommendations - Routes ranked by tailwind advantage on the homeward leg. Confidence levels from "ride either way" to "strong tailwind home."
- Segment colouring - Map view colours each route segment by wind condition (green = tailwind, red = headwind).
- Mobile-first - Designed for checking on your phone and sharing links with other riders.
- Route analysis - Automatically classifies routes as loops, out-and-backs, or point-to-point. Detects clockwise orientation. Simulates the ride segment by segment for wind scoring.
- Cafe stops - Routes show their cafe stop pinned at the right mile, geocoded via OpenStreetMap. Because the mid-ride coffee & cake is non-negotiable :).
- Road closure link - Links out to one.network to look for road closures. Ideally this would pre-analyse the route but this API is not publicly available, will aim to improve the experience.
- LBRCC club rides - Dedicated section for weekly group rides with wind analysis per ride, announcements, and group labels (G1/G2/G3).
- Strava integration - Paste any public Strava route URL to check today's wind conditions against it. (Currently hidden from the live site due to Strava's non-verified API limitations.)
- Community submissions - Submit routes via Strava URL. Moderated before appearing in the feed.
- Departure time picker - Check conditions for different times of day using forecast data.
- Dark mode - System preference detection with manual toggle.
- Framework: Next.js 16 with React 19 (App Router, TypeScript)
- Database: PostgreSQL on Neon, managed with Drizzle ORM
- Maps: Mapbox GL for interactive route visualisation
- Weather: Open-Meteo API (free, no key required) with 10-minute server-side cache
- Auth: Strava OAuth for route fetching
- Email: Resend for admin notifications on new submissions
- Styling: Tailwind CSS 4 with OKLCH colour system, Outfit + Geist fonts
- Hosting: Vercel
Tailwise simulates riding the route in both directions:
- The hourly wind forecast comes from Open-Meteo
- The ride is walked segment by segment at an assumed 16 mph, giving each segment a timestamp
- Each segment's tailwind component is scored against the forecast for the hour the rider reaches it. The ride home is scored on the wind at that time, not on conditions at departure
- Segments late in the ride are weighted more heavily, because a tailwind matters most on tired legs
- The direction with the better score wins. Thresholds: under 2 mph of advantage = "ride either way", 2-5 mph = "moderate", over 5 mph = "strong"
On a closed loop in steady wind, direction makes no difference: each stretch of tailwind in one direction becomes headwind in the other, and the two scores cancel. Direction advice matters when the wind shifts or builds during the ride.
The map's segment colouring comes from the same simulation, so the map and the recommendation always match.
# Install dependencies
npm install
# Set up environment variables (see .env.example)
cp .env.example .env.local
# Run database migrations
npx drizzle-kit push
# Start dev server
npm run devRequires: Node.js, a PostgreSQL database (Neon free tier works), Strava API credentials, Resend API key and a Mapbox token.
src/
app/ - Next.js pages and API routes
components/ - UI components (route cards, maps, forms, wind compass)
lib/
wind-advisor.ts - Core recommendation algorithm
route-analyzer.ts - Route classification and geometry
strava.ts - Strava API client with token caching
weather-server.ts - Weather fetching with in-memory cache
db/ - Drizzle schema, queries, migrations
constants.ts - Thresholds, API config, algorithm parameters
- Multi-club / multi-location - Currently built around Leighton Buzzard. The core logic is location-agnostic, it just needs a set of routes in a local area as seed data.
- Personal Strava library - Deeper Strava integration to analyse a rider's own saved routes, not just the curated set.
Live and being used by LBRCC club members. Actively maintained.
Built by Aaron Cottrell