Nitrite's built-in stop-word based tokenization is not a good fit for the title field of an article, but the ability to FTS article titles is valuable nonetheless.
Consider having two fields, a normal text field for the title that can match exactly and a hidden(ish) tokenized field.
Look into whether Nitrite allows full matching on FTS fields (probably not if memory serves me right).
Nitrite's built-in stop-word based tokenization is not a good fit for the title field of an article, but the ability to FTS article titles is valuable nonetheless. Consider having two fields, a normal text field for the title that can match exactly and a hidden(ish) tokenized field. Look into whether Nitrite allows full matching on FTS fields (probably not if memory serves me right).