Phonetic Name Matcher

Clusters

No likely duplicates at 90% similarity. Lower the threshold to catch more.

About this tool

Double Metaphone is Lawrence Philips' 2000 phonetic algorithm. It encodes how a name sounds rather than how it is spelled, and returns two codes: a primary for the most likely pronunciation and a secondary for a plausible alternative - which is why international names match so much better with it than with plain Soundex. Smith and Smyth both reduce to SM0 because the "th" digraph maps to 0 (the theta sound) and "i" versus "y" collapse to the same vowel code, so the two spellings cluster together. This implementation follows Philips' published C reference, with the initial "X" special case that yields SF/SFR for Xavier.

Soundex is computed alongside using the American Soundex rules (first letter kept, vowels and Y dropped while H and W stay transparent, adjacent duplicate codes collapsed, padded to four characters). Groups whose Soundex matches but whose Double Metaphone codes differ are flagged as weaker matches instead of being merged. Similarity scores use Levenshtein distance normalised to a percentage, plus Jaro-Winkler.

All of it - parsing, phonetic encoding, clustering, similarity, CSV parsing and export - runs in your browser; no name is ever uploaded. Input is capped at 2,000 names and the code table renders at most 400 rows, but clustering, dedupe and export still cover every name you pasted.