Semantic Similarity Calculator
Cosine, Jaccard, Levenshtein between any 2 texts
Similarity metrics
📚 Learn more — how it works, FAQ & guide Click to expand
Learn more — how it works, FAQ & guide
Click to expand
Semantic similarity calculator
Lexical + structural similarity between any two texts. Jaccard, cosine, Levenshtein.
How to use this tool
- 1
Paste two texts
Any length. Sentences, paragraphs, documents.
- 2
See similarity metrics
Jaccard, cosine, Levenshtein, length diff.
- 3
Pick interpretation
What each metric means for your use case.
Frequently Asked Questions
Which similarity metric to use?
Jaccard: word overlap only. Cosine: weighted by frequency. Levenshtein: character-level edit distance (catches typos). For RAG chunk dedup: cosine. For near-duplicate detection: Levenshtein ratio.
Is this real embedding similarity?
No — this is lexical similarity (client-side, no API). For true semantic similarity, you need embedding model. But for many cases (chunk overlap, cache key dedup, approximate match), lexical is fast + good enough.
You might also like
🔒
100% Privacy. This tool runs entirely in your browser. Your data is never uploaded to any server.