Text Similarity Checker
Compare two texts and measure similarity using three algorithms: Levenshtein distance, Jaccard coefficient, and LCS.
Samples
Text A Characters: 0Text B Characters: 0Levenshtein Distance: 0
Text Similarity Checker
This tool measures how similar two texts are using three complementary algorithms. Levenshtein similarity measures character-level edit distance — useful for detecting typos or near-duplicate strings. Jaccard similarity compares word sets — good for detecting paraphrasing. LCS (Longest Common Subsequence) measures shared character order — effective for detecting reordered content. No text leaves your browser.
Built and maintained by Meet Shah · Last updated
What this tool is used for
- Measuring how close two drafts are before deciding whether one is a rewrite.
- Checking whether two descriptions are near-duplicates.
- Comparing a submission against a source for overlap.
- Quantifying how much a document changed between versions.
- Choosing between two similarity measures for a specific comparison.
Frequently Asked Questions
- What does the similarity percentage measure?
- Jaccard similarity over word sets: the size of the intersection divided by the size of the union. It ignores word ORDER and word frequency entirely, so two documents with the same vocabulary in a different arrangement score identically.
- Why does it behave oddly on Chinese or Japanese text?
- Because those scripts do not put spaces between words, so naive whitespace tokenisation sees an entire document as ONE token — making the score degenerate to 0 or 100 with nothing in between. Correct handling needs dictionary segmentation.
- How is this different from Levenshtein distance?
- Levenshtein counts character edits and so is sensitive to order and to small typos; Jaccard compares vocabulary sets and is not. Use edit distance for near-identical short strings, and set similarity for documents.
- Does it detect paraphrasing?
- No. Substituting synonyms changes the word set, so a fully paraphrased passage can score near zero while saying the same thing. Semantic similarity needs embeddings, which is a fundamentally different technique.
- Is this a plagiarism checker?
- Not on its own — it compares two texts you supply, with no corpus and no web index. It tells you how much two documents overlap, which is useful for versions and drafts, not for detecting copying from unknown sources.
- How does document length affect the score?
- Substantially. A short text shares fewer distinct tokens, so a single common word moves the coefficient a long way — two three-word phrases can score 33% on one shared article. Scores are only comparable between texts of similar length.
- Should the comparison ignore stop words?
- It depends on what you are measuring. Keeping them inflates similarity for any two texts in the same language; removing them focuses on content, but also removes the function-word patterns that authorship analysis relies on.
Common errors and gotchas
- Reading a similarity score as evidence of plagiarism, which needs judgement rather than a number.
- Comparing texts of very different lengths, where most measures behave unintuitively.
- Using edit distance on long documents, where the cost is quadratic and the number is hard to interpret.
- Assuming the measures agree, since Jaccard and Levenshtein answer different questions.
- Comparing formatted against unformatted text, where whitespace dominates the difference.
Related Text Tools tools
Case Converter
Convert between Sentence, Title, camelCase, snake_case, and more.
Word Counter
Count words, characters, and analyze keyword density.
Text to Binary
Convert text into its 8-bit binary representation.
Binary to Text
Decode 8-bit binary back into readable text.
Text Reverser
Reverse text by characters, words, or lines.
Remove Duplicate Lines
Delete repeated lines and keep your list unique.
Remove Empty Lines
Strip blank and whitespace-only lines from text.
Sort Lines Alphabetically
Sort lines A–Z or Z–A, case-sensitive or not.