Spelling tolerance in search Pro
A search returning no results retries against the closest spelling in your titles and categories, and tells the reader which word it actually searched for.
Updated August 29, 2026
Someone types safeguardng and gets no results. The document is right there.
Switch on Documents → Settings → Search → Spelling tolerance and a search returning no results is retried against the closest spelling.
It is only ever a fallback
A search that found something is never second-guessed. <!– claim-ok: verified by the integration suite — "a search that found something reports no correction" –>Correction runs only when the exact search returned no results at all, so turning this on cannot change any result you already get today.
When a correction is used, the library says which spelling it searched for. It does not silently substitute — a search that quietly answers a different question is worse than one that returns no results.
Where corrections come from
Your document titles, categories and tags. Deliberately not the text inside your files.
A vocabulary drawn from the full text of a library is enormous and mostly noise, and correcting a misspelling towards a word that appears once inside a PDF produces a confident wrong answer. Titles and category names are what people actually type.
How conservative it is
- Words shorter than four characters are never corrected <!– claim-ok: verified by the integration suite — "a very short word is never corrected" –>. "cat" must not quietly become "car".
- The first letter has to match.
- One edit for a short word, two for a longer one.
If no candidate is close enough by those rules, the search returns no results, as before.
What still gets logged
The original query is what appears in your search log and download log, even when a correction rescued it. Recording the correction instead would erase the evidence that people are misspelling something — which is often worth knowing.
Still stuck? Email support@getfiledeck.com.