Word ladder

Connect two English dictionary words of the same length by replacing exactly one letter at a time. Every intermediate word must be in the same dictionary. Words have 1–12 letters; outer spaces and case are ignored. Digits, punctuation, accents and internal spaces are rejected rather than silently removed. Both endpoints must be known, even when they are identical.

The result is one shortest chain in this dictionary, not every possible chain. The number of changes is the number of links, one less than the number of words. Identical known endpoints need zero changes. Highlighted letters differ from the preceding word. Copy and TXT contain the whole chain, one word per line, without headings or timing data.

The general English word list includes uncommon/historical entries, names and abbreviations. A shortest chain may therefore use unfamiliar words and may not be accepted by another game’s dictionary. Other interface languages do not change the vocabulary. The search does not rank words by familiarity or usage frequency.

Only words of the required length are searched. A breadth-first search expands complete levels from both endpoints, generating one-letter replacements. Tied shortest chains are resolved deterministically. Large searches yield regularly and show examined-word counts. The limit is 100,000 stored search entries across both directions: reaching it reports a limit, not proof of disconnection. If an entire search frontier is exhausted without meeting, no chain exists in this dictionary.

The word list is loaded on demand. Editing either endpoint removes old results and obsolete requests cannot replace them. Failed loads can retry; a failed copy leaves TXT available. Inputs stay in memory across tool navigation; reloading restores cold and warm. All search work runs in the browser without uploading the words.