Loading page
An anagram solver rearranges a set of letters into words that use every one of them. This one searches the 77,636 single-word entries in AltF Lexicon, matches your letters against 72,807 distinct letter sets, and returns every real word that spends the whole rack. Type ? for a blank tile.
Type at least two letters and the answers appear here.
Every word in the dictionary has a signature: its letters, sorted. LISTEN, SILENT, ENLIST and TINSEL all reduce to eilnst. Sorting throws away the arrangement and keeps only the multiset of letters, which is exactly the thing anagrams have in common.
The solver builds that signature for all 77,636 single-word entries once, and groups them into 72,807 buckets. When you type letters, it sorts them the same way and looks the bucket up directly — one hash lookup, not a scan of the dictionary. That is why the answers arrive while you are still typing.
A blank breaks the shortcut, because ? could be any of twenty-six letters and there is no single bucket to look in. With blanks the solver walks every bucket of the right length and asks whether your letters cover it with at most that many gaps. It is a full pass over the index rather than one lookup, which is why blanks are capped at two: a third would widen the answer set past the point where it answers anything.
The index holds the 77,636 AltF Lexicon entries whose headword is a single unbroken run of a–z, between 1 and 31 letters. Phrases (q fever), hyphenated compounds and anything carrying a digit or an apostrophe are excluded, because a rack of tiles cannot produce them.
Answers are ordered by length, then by how common the word is, then alphabetically. The five-segment meter beside each answer is that commonness band, measured against a corpus of everyday English rather than assigned by an editor — it is the quickest way to see that silent and drepanis are not equally useful to you.
The word list comes from Princeton University’s WordNet, with commonness bands derived from a frequency corpus of everyday English. Definitions on the linked entry pages are WordNet’s; pronunciation comes from the CMU Pronouncing Dictionary. Full sources and licences.