They actually hand wave and gloss over two of the other biggest drawbacks:
- Starlight is 6 years less mature than docusaurus
- the people who maintain starlight have some level of billions less at their disposal to keep the project going. If Astro and the company behind it go belly up now you have a new problem on your hands
The main appeal I can see is for someone who wants a lot more extensibility and control over the design aspect of their docs. For those who just need to slap some pretty good looking docs into a well supported framework for the next 10 years would do better with the more battle tested framework supported by BigCorp
> the people who maintain starlight have some level of billions less at their disposal to keep the project going. If Astro and the company behind it go belly up now you have a new problem on your hands
Starlight and Docusaurus aren't much different in this aspect. Meta could decide to stop paying slorber (the main maintainer of Docusaurus) at any time. I hear from friends in Meta that Docusaurus isn't even widely used internally.
Both projects are MIT-licensed, at least, and that's a plus for continuity if someone ever wants to fork them.
Disclosure: Astro core team
Yes, the difference here is that Astro considers Starlight to be a core part of our project. We use it for our own docs, as well as it being the basis for a significant percentage of our users' sites. I've no reason to believe Docusaurus's funding is in any danger, but I think the Open Collective that pays for full time Starlight development is probably a more dependable source of funding than Meta.