Super interesting question! Searches are always tricky to implement, because of the sheer amount of data available.
Having had a look at the documentation, Ghost only uses the title and excerpts – and only builds the search index on the last 10,000 posts. In my opinion, a pretty good compromise between speed and depth of the index.
To give you some context, implementing a full-text search across all content can be quite resource-intensive. Factors like indexing every word, handling synonym recognition, and ranking results can quickly become super complex.
As a reference, in my day job, our full-text search for around 10,000 articles requires a dedicated virtual server equipped with 8GB memory and a powerful CPU.
Considering that Ghost aims to remain lightweight and accessible, focusing the search on titles and excerpts is a sensible decision, I think. It ensures efficiency without overburdening the infrastructure of small blogs