Document Similarity Checker
Compare two documents for overlapping passages.
Loading…
About Document Similarity Checker
Two drafts, two contract versions, or a rewrite and its source: the question is how much wording they share. Paste each text into its own box and the checker lowercases it, splits it into words and builds every run of three consecutive words. The score is the share of distinct phrases that appear in both texts, and the matching phrases are listed so you can see what overlaps. It measures wording, not meaning.
Features
- Jaccard score over distinct three-word phrases
- Lists the matching phrases with the total number found
- Ignores case and punctuation
- Word splitting that copes with unspaced scripts such as Chinese and Japanese where the browser supports it
- Falls back to word-by-word comparison when a text is under three words
Limitations
- It measures shared wording only. A paraphrase with different words scores low even when the meaning is the same, and the score is not a plagiarism verdict or an AI-authorship check.
- Each text can hold up to 200,000 characters, and only the first twelve matching phrases are listed, although the total is shown.
- Phrases are counted once per text and their order is ignored, so repeating a sentence or moving whole paragraphs around barely changes the score.
How it works
- Paste the first text into Document A and the second into Document B.
- Check the character counter under each box, which counts up to 200,000 characters.
- Select Compare documents.
- Read the percentage under Lexical overlap and the bar below it, which changes color as the score rises.
- Go through the matching phrases to see exactly which wording the two texts share.
- Use Reset to clear both boxes and start another comparison.
Example input
Document A: "The quick brown fox jumps over the lazy dog."
Document B: "A quick brown fox jumped over a lazy dog."Frequently asked questions
- It is the number of three-word phrases found in both texts divided by the number found in either one. The pair "The quick brown fox jumps over the lazy dog." and "A quick brown fox jumped over a lazy dog." share only "quick brown fox", which is one phrase out of 13 distinct ones, so the score is 7.7%.
- No. It compares two texts you supply and has no model of how AI writes. Two human-written passages that quote the same source can score high, and AI text that shares no wording with the other document scores near zero.
- Because it compares words, not ideas. Synonyms, a changed word order or an extra adjective break up the three-word phrases, so a careful paraphrase of a paragraph can overlap very little. For that kind of check you need a semantic comparison, which this page does not do.
- If either text has fewer than three words, both are compared word by word and the page says so. Words are split with the browser's own word segmentation where available, which also finds word boundaries in Chinese, Japanese and Thai, and text is Unicode-normalized before it is compared.
- No. The comparison is a script that runs on the page, there is no upload step, and nothing is stored. Select Reset to clear both boxes. The Share button copies only the tool's address, never the text you pasted.
Related tools
- Advanced Similarity CheckerCompare two texts by lexical overlap and semantic meaning.
- Plagiarism CheckerCompare two documents for potentially matching content.
- Word CounterCount words, characters, sentences and get reading time.
- Compare PDFCompare two PDFs and see what changed.
- LLM Token Analyzer & Prompt OptimizerCount tokens, estimate API usage, optimize prompts, and discover ways to reduce unnecessary token consumption.