Astro is a web frameworks geared towards sites that are mostly statically generated. One of the main use cases for this is blogs, and in fact its what this site is built in.
Astro ticks a lot of boxes for me. Large Frameworks leave me feeling far from the final product shipped to users, which is dissatisfying in a personal project. Astro feels much more like pure html, while still offering templating and the ability to dynamically generate pages. Writing to code to generate a page, rather than code to control one, feels far better for a project like this.
Its also extremely fast to get off the ground with. Astro offers a very robust tool chain including tools for starting a new project using a template (yarn create astro), a tool for installing add-ons with default settings (astro add), and an option to render markdown files as html within templates via file based routing, so no need to configure that if you just want to start showing html to users. I was able to get a project with tailwind running with custom metadata included in the content files within an hour which feels great for a personal project.
And while I’m on the topic, the ability to write articles as markdown and include data in front matter that can then be accessed within your code is wonderful. While this isn’t unique to Astro, this is one of the things Astro does that fits the File over App philosophy very well, which is a huge bonus for me since this is ultimately a writing project first and foremost. The others are the way it encourages relatively vanilla JS, without reliance on platform-specific hooks and data types (Zod is used to cast data from article front matter to standard TS types), and the way that interactive components can be written in a variety of other frameworks via one of a number of easily installed adaptors. If I build and interactive elements I’ll likely do it in Vue, and that will be code I can take with me along with the content itself if I ever migrate away.
The only downside to Astro really are the lesser (but still present) support for on demand rendered pages and it’s relatively small user-base. But Astro is very specifically targeted at static sites meaning you’re pregnant better off with another choice for anything dynamic anyway, and 9% of the static site market is still tens of thousands of users, plus with recent dev satisfaction survey wins it’s only lively to grow in the near future1.
Overall I’m very satisfied with Astro, and I intend to continue using it for the foreseeable future. I’d also like to use it for other projects as and when they come up, provided that they don’t involve too much dynamic content.