Duplicate Word Remover

Remove repeated words from your text

A duplicate word is a word that appears more than once where only one instance was intended - most often the accidental "the the" that survives an edit, but also every repeated tag in a keyword list. This tool strips those repeats in one pass, using whichever definition of "duplicate" matches your situation.

The doubled-word typo is unusually hard to catch by eye. Reading is predictive: the brain supplies expected function words before the eye reaches them, so a repeated the or and is often skipped entirely - the well-known "Paris in the the spring" illusion. Spellcheckers are little help either, because both copies are correctly spelled; Microsoft Word and Google Docs flag consecutive repeats specifically as a separate grammar rule, and plain text editors do not flag them at all. Automatic transcription introduces its own version, since Whisper and similar models faithfully reproduce the stutters and restarts of natural speech.

Two modes are available to suit different needs:

• Consecutive duplicates: removes only immediately adjacent repeats. "the the cat sat on the mat" → "the cat sat on the mat". The second "the" later in the sentence is kept because it is not consecutive.
• All duplicate words: keeps only the first occurrence of each unique word across the text. Useful for keyword deduplication or list cleaning.

Common use cases:

• Transcription cleanup: fix speech-to-text "the the" errors
• Writing proofreading: catch accidental word repetition during editing
• Keyword lists: deduplicate tag or keyword lists
• Data cleaning: remove word-level duplicates from structured text

Choosing the right mode - and reviewing the result

Consecutive mode is the safe one and is what you want for prose. It only touches adjacent repeats, so legitimate repetition elsewhere in the sentence is untouched. All-duplicates mode is deliberately aggressive: it is built for lists, tag fields, and keyword sets, and it will happily destroy ordinary writing, because English sentences reuse the, of, and a constantly by design. To limit the damage, all-duplicates mode operates line by line rather than across the whole document, so an item on line 4 does not suppress the same word on line 9.

Comparison is case-insensitive, so "The the" collapses to a single The - the first occurrence keeps its original capitalisation. The two modes differ on punctuation: consecutive mode requires the repeats to be separated by nothing but whitespace, so New York, New York survives the comma intact, while all-duplicates mode strips punctuation before comparing and would treat both York tokens as the same word. Either way, a genuine repeat such as had had in the past perfect or that that in a relative clause will be collapsed. Skim the output before accepting it rather than pasting it straight back into a document.

For removing duplicate lines rather than words, see our Remove Duplicate Lines tool; for collapsing runs of whitespace left behind after an edit, see Remove Extra Spaces. Deduplication runs entirely as local JavaScript, so transcripts and unpublished drafts never leave your device.