Typo tolerance is a search engine's ability to return correct results despite misspellings, transpositions and missing characters in the query.
Also known as: fuzzy matching
It is usually implemented with edit distance — how many single-character changes separate two strings — scaled by word length, so short words allow one correction and longer words allow two. Phonetic matching and keyboard-adjacency weighting refine it further.
Tolerance must be bounded. Applied too aggressively it destroys precision on identifiers, where one character genuinely distinguishes two different products. The usual compromise is strict matching on codes and numeric tokens, fuzzy matching on natural-language words.
A misspelling returning nothing is, from the visitor's side, indistinguishable from the site not stocking the item. That is why typo tolerance is a stock-availability problem disguised as a text problem.