Free tool
Word Repetition Checker
Paste your text and instantly see which words repeat too close together. Built on the same proximity engine as our manuscript editor — a word is only flagged when its next occurrence appears within 250 characters. That's proximity analysis, not a naive global search.
0 / 50,000 · Up to 50,000 characters. The repetition analysis runs locally in your browser.
Repetitions
Paste your text above — repetitions are flagged as soon as two occurrences of a word appear within 250 characters of each other.
How proximity repetition detection works
Authors repeat words on purpose all the time — often without noticing. A simple word counter can't help, because a word repeated twenty pages apart is usually fine. That's why our checker flags a repetition only when two consecutive occurrences of the same word sit within 250 characters of each other: close enough to catch the reader's eye and pull them out of the story. Stop words (and, the, der, die, …) are ignored automatically, and the analysis runs entirely in your browser.
Up to 50,000 characters. The repetition analysis runs locally in your browser.
The repetition analysis itself runs locally in your browser — nothing is uploaded. Only when you request an AI synonym suggestion is a short excerpt (about 100 characters around the word) sent to generate context-aware synonyms.
Repetition or style device?
Not every repetition is a mistake. When a word returns within a few sentences, it reads like an echo — readers stumble, even if they can't say why. But some repetitions are deliberate: anaphora, leitmotifs, a rhythm you build on purpose. The difference is usually proximity and intent. A motif that returns after chapters carries the story; the same word three times in one paragraph distracts from it.
She looked at him. She looked at the rain. She looked at nothing.
The same sentence opening three times is anaphora — deliberate rhythm when it's meant. The checker flags it so you can decide, not so you have to change it.
The words that repeat most often
In English manuscripts, verbs of seeing and moving, body parts, and adverbs like "suddenly" pile up quickly. Short words and stop words such as "and", "the" or "just" are ignored automatically — including the dialogue tag "said". Content words are the ones worth a second look.
Fixing flagged repetitions: four techniques
Once a word is flagged, there are four honest options. Replacing it mechanically is the weakest — the others almost always read better.
01
Pick a fitting synonym
A synonym only helps if it fits the scene. Geist suggests context-aware alternatives directly in the text — no blind thesaurus dive.
02
Rebuild the sentence
Instead of swapping the word, change the construction: "She looked at him" becomes "Her gaze met his".
03
Merge or compress
Two sentences that say almost the same thing often collapse into one: "He walked to the door. He walked slowly." becomes "He walked slowly to the door."
04
Cut it
Often the second mention is simply redundant. Deleting is the cleanest fix — what remains reads more confident.
Checking a whole manuscript
The checker is built for real chapters, not just snippets. Four steps take you from paste to polished:
- 1
Paste chapter by chapter — up to 50,000 characters at a time, enough for long chapters or a short story.
- 2
Review the flagged words: every repetition appears as a chip, and one click jumps to its first occurrence in the text.
- 3
Smooth individual spots: clicking a highlighted word opens dictionary and AI synonym suggestions you can apply in place.
- 4
WriteGeist runs the same proximity engine live while you write, so repetitions never survive to the final draft.
Frequently asked questions
What counts as a word repetition?
A repetition is flagged when the same word appears at least twice and two consecutive occurrences are closer than 250 characters apart. Words repeated far apart are not flagged — most of the time they are intentional.
Why does proximity matter?
A word that repeats within a short distance is noticeable and reads as echoing. But a repeated motif that only returns after pages is usually deliberate. Proximity detection tells the two apart automatically.
Does the checker work for German and English?
Yes. The engine handles German and English with a shared stop-word list, and synonym suggestions are available in both languages. Auto-detection follows the language of the page you're on.
Do I need an account to use it?
No. Paste your text and the proximity analysis runs immediately for free. You get 10 AI synonym suggestions per browser session without an account — creating a free WriteGeist account unlocks unlimited proximity checks and the full manuscript editor.
Is my text uploaded anywhere?
The repetition analysis itself runs locally in your browser — nothing is uploaded. Only when you request an AI synonym suggestion is a short excerpt (about 100 characters around the word) sent to generate context-aware synonyms.
Do "walk" and "walked" count as the same word?
No. The checker compares word forms exactly (case-insensitive): "walk" and "walked" are counted separately, as are "eye" and "eyes". Inflected variants that still read as a repetition won't always be highlighted — keep an eye out for those.
Are dialogue tags like "said" flagged?
"Said" is on the shared stop list and is ignored. German tags like "sagte" are not, so they can surface — and rightly so: if every other line is someone saying something, the dialogue turns monotone. Vary tags sparingly instead of replacing them mechanically.
Why 250 characters as the threshold?
250 characters is roughly a short paragraph — the distance at which readers consciously notice an echo. Words repeated across longer stretches are usually intentional and are not flagged.
Does it detect repeated sentence openings?
Not directly — the checker works word by word. A repeated opening like "She looked… She looked…" still surfaces because the repeated word itself is flagged. For rhythm and sentence structure, read the marked passages aloud; your ear catches what a list can't.
Why are character names flagged?
Names repeat in every scene — and that's fine. The checker doesn't know your cast, so "Lyra" or "Mors" can appear as chips. Treat name chips as expected and focus on the words around them.
Is the checker useful for editing and translation?
Yes. Editors and translators use the proximity list as a fast quality check on revised or machine-translated chapters, where phrasing from the source text often creates repetitions. Paste the chapter, walk the chips, smooth the spots.
Free tool
Like this clean repetition analysis?
WriteGeist gives you live proximity checks, character wikis, and timelines directly inside a full chapter-based manuscript editor.
No credit card required