Text · Made a little simpler

Text Similarity Checker

Compare two texts with cosine and Jaccard similarity. Inspect shared and exclusive vocabulary, check frequencies and download the report locally.

Free to useNo account neededProcessed on your device
Your workspaceProcessed on your device

Add this tool to your website

Paste this code into your page to add the tool and a visible ToolMellow credit link.

HTML embed code
Preview embed

Text Similarity Checker: how to use it

  1. Paste two passages or open UTF-8 text files into Text A and Text B.
  2. Choose whether to distinguish uppercase and lowercase, then select Compare texts.
  3. Compare cosine and Jaccard similarity and filter the vocabulary table. Download the JSON report to keep all terms and counts.

Limits and notes

Use up to 50,000 UTF-16 units per text. Tokenization uses your browser’s Unicode word boundaries. NFC normalization and straight versus curly apostrophe equivalence apply in both case modes. The scores use the supplied texts and count words without taking their order into account.

How your privacy works →

Text Similarity Checker: common questions

Why can the cosine and Jaccard scores be different?

Cosine uses how often each word appears. Jaccard counts each distinct word once and divides the number shared by the number that appear in either text.

Do accents affect the vocabulary match?

Accents remain significant. A single é and an e followed by a combining acute accent match because they normalize to the same Unicode form.

What happens when only one passage has words?

Jaccard similarity is zero. Cosine similarity requires words in both passages, so it is unavailable.

Guides