Soundex Calculator

How to use

1. 1Pick Encode word, Compare two words, or Batch list at the top.
2. 2In Encode word mode, type a name or word into the input. The Soundex, Refined Soundex, Metaphone, and NYSIIS codes update on every keystroke.
3. 3In Compare two words mode, type both spellings. Each algorithm reports Match or Different and shows the two codes side by side.
4. 4In Batch list mode, paste one word per line. The table shows every algorithm for every line. Click Copy as CSV to paste the result straight into a spreadsheet.
5. 5Use the Which algorithm should I use? card at the bottom to decide between Soundex (genealogy), Refined Soundex (lower collisions), Metaphone (general English), or NYSIIS (surname-heavy data).

About this tool

Soundex Calculator encodes any name or word with the four most widely used phonetic algorithms and shows the results side by side so you can pick whichever one fits your data. Soundex (Russell and Odell, 1918) is the classic four-character code still used by the US Census Bureau, the National Archives, and most genealogy software; it keeps the first letter and turns the remaining consonants into three digits using six well-known groups. Refined Soundex splits those six groups into nine to reduce collisions and is used by some database engines including MySQL refined Soundex. Metaphone (Lawrence Philips, 1990) is a more accurate encoder that handles the English letter combinations Soundex misses, including ph, gh, kn, sch, and th; its codes are variable length and tend to read closer to the actual pronunciation. NYSIIS (the New York State Identification and Intelligence System, 1970) produces a six-character key with explicit handling for the common name prefixes Mac, Mc, Kn, Pf, Sch and for the trailing -ed, -nt, -nd, -rd, -rt, -dt endings; it is still used by US law enforcement and government databases. The tool offers three modes: Encode word shows all four codes for a single input, Compare two words tells you which algorithms consider the pair a match, and Batch list takes one term per line and outputs a table you can copy as CSV into a spreadsheet. Letters only are encoded: diacritics are stripped, digits and punctuation are removed, and hyphens are ignored so names like Anne-Marie and surnames with accents work cleanly. Useful for genealogy and family-tree research, library catalog reconciliation, voter-roll de-duplication, healthcare record linkage, customer-database fuzzy lookup, ETL pipelines that match free-text names, and any application that needs to find records that sound alike even when they are spelled differently. Pair phonetic matching with Levenshtein or Jaro-Winkler distance from the String Similarity Checker for the strongest results; phonetic codes alone are intentionally coarse and will produce occasional collisions on unrelated words. Every algorithm runs entirely in the browser. Names you paste are never uploaded.

Free to use. Works in your browser. No signup, no login.