Typesense works by determining how similar two terms are based on how many edits are needed to turn one into another - typically called "edit distance" or "Fuzzy matching" - this is a mathamatical process and does not distinguish between "real" and "not real" words.