When a text field is indexed (and when a match query runs), an analyzer turns text into tokens: Character filters — e.g. strip HTML. Tokenizer — split into words (standard, whitespace, ngram, edgengram). Token filters — lowercase, stop, stemmer, synonym, asciifolding. Built-in analyzer Result for "The Running Foxes!" standard the, running, foxes english run, fox (stop words removed, stemmed) keyword The Running Foxes! (one token) whitespace The, Running, Foxes! Use analyze to see exactly what is indexed — most "why doesn't this match?" questions are answered there. Autocomplete typically uses an edgengram analyzer at index time and standard at search time.