Fuzzy duplicate finder

Exact duplicate detection misses the most common real-world case — the same customer entered as “Jon Smith” and “Jonathan Smith,” or “123 Main St” versus “123 Main Street,” where the records refer to the same person or thing but don’t match character-for-character. The Fuzzy Duplicate Row Matcher compares rows using similarity scoring rather than exact matching, flagging likely duplicates even when names, addresses, or other fields are spelled or formatted slightly differently, so you can review and merge them rather than missing them entirely.

The kind of data this runs on — customer records, contact lists, patient or client rosters — is usually sensitive enough on its own, and deduplication specifically involves comparing records against each other in a way that could reveal overlaps a person might not want surfaced (two records under different names turning out to be the same individual, for instance). All the comparison and scoring happens in your browser, so the dataset being deduplicated never has to be uploaded anywhere to get matched.

Upload your file and choose which columns to compare (name and address are common choices, but any text column works), then set a similarity threshold — higher thresholds catch only very close matches, while lower thresholds catch more potential duplicates at the cost of more false positives to review manually. The tool presents matched pairs side by side with a similarity score, and you decide whether to merge, keep both, or dismiss each flagged pair rather than having any merge applied automatically.

FAQ

Why did it flag two rows as duplicates that I know are actually different people?

This happens with common names or shared addresses (family members at the same address, for instance) — a high similarity score reflects textual closeness, not certainty, which is exactly why the tool surfaces matches for manual review rather than merging automatically.

What similarity threshold should I start with?

A moderate threshold that catches obvious variations (typos, abbreviations, minor formatting differences) without flagging too many unrelated rows is usually the right starting point — review the first batch of results and adjust up or down depending on how many false positives or missed matches you’re seeing.

Can I compare across multiple columns at once, like both name and address together?

Yes — combining columns generally improves match accuracy, since two records with a similar name AND a similar address are much more likely to be true duplicates than a match on either field alone.

Once duplicates are resolved, this pairs naturally with Merge CSVs if you’re consolidating multiple source files, and with CSV cleaner if the underlying formatting inconsistencies (casing, whitespace) are contributing to the fuzzy matches in the first place.