I run a surf forecast for 20 breaks in Morocco on EUR 0/month. Here's the stack.

I operate a surf forecast service for 20 breaks along Morocco's Taghazout coast, a stretch of Atlantic coastline between Agadir and Imsouane. Previously, surf enthusiasts had to rely on three global forecast sites, none of which accurately distinguished between individual breaks. To address this, I created taghazout.io, a platform covering 20 named breaks, available in 10 languages, and operates at zero cost per month.

The platform's architecture is straightforward, utilizing hand-rolled PHP with no frameworks, build steps, or dependencies. The server-rendered pages are optimized for users accessing the information via mobile phones or cafe Wi-Fi connections, often on slow 3G connections, making time-to-content the primary metric. The forecast data is sourced from two ocean models: Open-Meteo, under a CC BY 4.0 license, and NOAA WaveWatch III via PacIOOS.

These models often disagree, with WaveWatch III sometimes providing readings up to 55% higher than Open-Meteo. To address this discrepancy, the platform computes agreement over a 72-hour window, displaying a confidence flag (agree, mixed, or diverge) to users. This information is crucial for surfers in shallow-reef setups, where coastal bathymetry can significantly impact wave conditions.

The platform also emphasizes the importance of honest data over impressive data, providing clear caveats for each derived number and its source. The tide readout, for example, specifies the sea level based on the low tide of the current window, rather than presenting it as a harbor gauge. All forecast data is publicly available as keyless JSON, encouraging transparency and collaboration.

The platform's source code is open-source, licensed under MIT, with Python and JavaScript clients available on GitHub.