If you're looking for a Dark Sky API replacement, Weather Machine is a good option. It combines weather data from several forecast APIs into a single interface, including a Dark Sky-compatible JSON endpoint and a GraphQL endpoint for conditional requests. It's also very efficient, with a sophisticated forecast-aware caching system that can dramatically reduce data usage, and is geared for developers and companies that want to build weather data into their apps without the hassle of managing multiple APIs.
Tomorrow.io is another all-purpose option that offers high-resolution weather data and forecasts. It includes real-time weather data, weather forecast API, historical weather API, and a generative AI for personalized recommendations. The service is geared for businesses that need to optimize operations and protect against weather risks, so it's good for supply chain and disaster recovery apps.
For hyper-local weather forecasts, Atmo offers AI-based systems that use deep learning models to make forecasts. It consumes global weather data from a variety of sources and produces forecasts at unprecedented speed and resolution, which is good for apps that need detailed weather information at very small scales.
Last, if you need raw model data from government agencies like NOAA, check out Pirate Weather. It supports multiple model sources and offers sub-hourly forecasts, short-term smoke forecasts, and extended hourly data. The API is customizable and can be used with tools like Home Assistant, so it's good for different weather-related needs.