[combatentropy]

42 MB for the table, which has columns for the address, title, plain-text body, and computed text vectors for 43,000 pages (web pages and office documents of average length). Then another 100 MB for the GIN index on the text-vector column.

[ngrilly]

That's only 977 bytes per page (42 MB / 43,000 pages). Are you sure about the numbers? Maybe the plain-text body is stored in a TOAST table?

Anyway, it looks like the whole dataset can fit in RAM, which explains the excellent performance, even with relevance ranking.

[combatentropy]

Whoops! Yes, there is a corresponding TOAST table that I had to track down.

  Table: 341 MB (main table + TOAST table)
  Index: 100 MB