Soundex Calculator
Generate the Soundex phonetic code for any word or name — free online Soundex algorithm calculator.
The Soundex code updates live as you type — enter multiple names, one per line, to compare their codes side by side.
Robert
Rupert
Smith
Smyth
How Soundex Works
Soundex is a phonetic algorithm developed in the early 1900s (and still an ANSI/NIST standard) that encodes a word — usually a name — into a 4-character code: one letter followed by three digits. Words that sound alike, even when spelled differently, often produce the same code, which is what makes Soundex useful for fuzzy, spelling-tolerant name matching.
The algorithm works in a few steps:
- Keep the first letter of the word exactly as it is.
- Convert every remaining consonant to a digit using this grouping: 1 = B, F, P, V; 2 = C, G, J, K, Q, S, X, Z; 3 = D, T; 4 = L; 5 = M, N; 6 = R.
- Drop all vowels (A, E, I, O, U) and Y — they don't get a digit.
- If two letters that map to the same digit sit next to each other, only the first is kept. H and W are treated as invisible when checking for this — so a repeated digit separated only by H or W is still merged into one, while a vowel in between keeps both digits.
- Pad the result with trailing zeros, or cut it short, so it's always exactly 4 characters.
Example
"Robert" → R163. R is kept as the first letter. O is a vowel and is dropped. B maps to 1. E is a vowel and is dropped. R maps to 6. T maps to 3. That gives R-1-6-3.
"Rupert" → R163. Despite the different spelling, the same steps produce the identical code — U, E are vowels (dropped), P maps to 1, R maps to 6, T maps to 3 — which is exactly why Soundex is used to catch alternate spellings of the same name.
"Smith" and "Smyth" both → S530. In "Smith", S is kept, M maps to 5, I is a vowel (dropped), T maps to 3, H is dropped, giving S-5-3, padded to S530. In "Smyth", S is kept, M maps to 5, Y is treated like a vowel (dropped), T maps to 3, H is dropped — the same S-5-3-0. Two completely different spellings, one identical code.
Common Use Cases
- Genealogy and ancestry research — historical records often misspell surnames, so genealogy databases (including US and UK census indexes) index names by Soundex code to find every spelling variant of a family name at once.
- Fuzzy database name matching — CRM and customer databases use Soundex to catch duplicate records for the same person entered with different spellings (e.g. "Catherine" vs "Kathryn").
- Spell-tolerant search — search tools use Soundex codes as a fallback so a misspelled name still surfaces the record the user was looking for.
- Data deduplication — grouping records by Soundex code is a quick way to flag probable duplicate names for manual review before merging datasets.
FAQs
What is Soundex?
Soundex is a phonetic algorithm that converts a word — typically a name — into a 4-character code (one letter plus three digits) representing how it sounds, so that similar-sounding words share the same code regardless of small spelling differences.
Why do two different spellings get the same code?
Soundex groups consonants that sound alike (like B, F, P, V) into the same digit and ignores vowels entirely, since vowels vary the most between alternate spellings of the same name. Two names that sound similar, such as "Smith" and "Smyth" or "Robert" and "Rupert", follow the same consonant pattern and so end up with the same code even though they're spelled differently.
Is Soundex case-sensitive?
No. Soundex first converts the input to uppercase, so "robert", "Robert", and "ROBERT" all produce the identical code, R163.
Does Soundex work for non-English names?
Soundex was designed around English-language pronunciation and spelling patterns, so it works best on names common in English-speaking records (which is why it's still widely used in US and UK genealogy archives). It can still be applied to non-English names since it only looks at Latin letters, but the codes it produces may not reflect how those names actually sound in their original language — algorithms like Double Metaphone or NYSIIS tend to handle non-English names more accurately.