Name Confusion Checker

Nobody notices at the naming stage, because you look at one name at a time. Readers see them together, and by then the names are on things. A cast with Marta and Marla in it, a team with Jon and Ron, a product range with Aurex and Aurix — this compares every pair in your list and tells you which ones will be misread, with the reason for each.

Your names

One per line. Everything stays in your browser — nothing is uploaded, stored or put in the address bar.

Paste a list of names and the pairs people are likely to mix up will appear here.

What this generator does

Compares every pair of names in your list and reports the ones likely to be confused, scored from 0 to 100 and sorted worst first. Each flagged pair says which tests fired: how few letter changes separate them, whether they share a phonetic key and so sound alike aloud, whether they start with the same letter, whether they rhyme, and whether they use identical consonants and differ only in vowels. Exact duplicates are reported separately, because that is a different problem with a different fix.

How to use this tool

Paste your names, one per line.
Press Check the list.
Read the pairs it flags, worst first — each one says why it was flagged.
Change the ones you agree with. The tool has an opinion, not the last word.

Understanding the controls

Your names: One per line, up to 200. Blank lines are ignored. Case does not matter for the comparison, and the names are shown back exactly as you typed them.

Worked examples

Marta / Marla: One letter apart, same first letter, same length — flagged high.
Jon / Ron: One letter apart and they rhyme, which is what catches this pair.
Sean / Shaun: Spelled differently, same phonetic key — a pair that only causes trouble aloud.
Priya / Wolfgang: Nothing in common. Not flagged, which is the other half of being useful.

Common use cases

Checking a cast of characters before a manuscript goes out
Testing a product or brand shortlist for names that will be confused in speech
Making sure a class list or team roster does not have two names that get mixed up
Naming variables, services or environments where a misread costs an outage
Checking a list of usernames or handles for near-collisions

How this generator works

Every pair is put through five tests and the results are combined into a score. Spelling closeness uses Levenshtein edit distance — the fewest single-character insertions, deletions or substitutions that turn one name into the other — divided by the length of the longer name, so that one letter different in three counts for much more than one in twelve. Sound uses Soundex, the standard phonetic key, which is deliberately crude but is defined precisely enough to be tested against its published reference values rather than being a set of rules only this tool agrees with. Rhyme compares the rime, meaning everything from the last vowel onwards, because a fixed number of trailing letters fails on exactly the names where rhyme matters most: comparing the last three letters of Jon and Ron compares the whole word. The remaining two tests look at the first letter, which is what the eye scans for in a list, and the consonant skeleton, which is most of what is left when a reader skims. A list of forty names is 780 comparisons and takes no measurable time.

Randomness and fairness

Nothing here is random. The same list always produces the same findings, which is what makes the result something you can check and argue with rather than reroll.

For how randomness is produced across the whole site, see how Generate Random works.

Assumptions this tool makes

Names are compared as written, in the Latin alphabet. Characters outside A-Z are ignored for the phonetic comparison.
Confusion is treated as symmetric — if A is easily read as B, B is easily read as A.

Limitations and good to know

It says nothing about trademarks, registrations, domain availability or whether a name is free to use. Those need a search of the appropriate register, and past a certain point a solicitor. Similarity of spelling is not legal clearance and must not be read as any part of it.
The scoring weights are a judgement, not a measurement. The output is a shortlist for a person to look at, not a number to act on directly.
Soundex was designed for English surnames. It is a poor guide to how names from other languages sound, and it will both miss real clashes and flag pairs that sound nothing alike outside English.
It compares names to each other, not to the world. Two names that are fine together may each clash with something already on the market.
It cannot know your context. Two similar names matter enormously on a ballot paper and not at all in a novel where the characters never share a scene.

Common mistakes

Reading a low score as a clearance: It means these two names are unlikely to be mixed up with each other. It says nothing about whether either is available, sensible, or already somebody else's.
Changing every flagged pair: The threshold is set to surface things worth a glance. Plenty of flagged pairs are fine in context — two similar names that never appear on the same page do no harm.
Only checking the names you invented: Paste the whole list, including the ones that were never in question. Clashes turn up between a new name and an old one far more often than between two new ones.

Practical tips

Read the flagged pairs aloud. The phonetic tests catch what the eye misses, and hearing them is the fastest way to decide whether the tool is right.
For fiction, vary the first letter and the number of syllables rather than the middle of the word. Readers navigate by shape and opening, not by spelling.
For anything spoken over a phone or a radio, weight the phonetic flags much more heavily than the spelling ones.

Troubleshooting

It flagged a pair that is obviously fine: Look at the reasons it gives. Usually one weak test fired — a shared first letter, say — and the score sits near the bottom of the range. Those are labelled 'worth a look' rather than 'very close' for exactly this reason.
It missed a pair I know gets confused: The tests are about spelling and English phonetics. Two names confused because they belong to similar characters, or because they mean the same thing, are beyond anything a string comparison can see.

Privacy and your data

Your names never leave the browser. They are not uploaded, not stored and never written into the page address — which matters here, because a list of candidate brand names or real people is often confidential. Analytics records only that the tool ran and how many pairs were flagged, never the names or the findings. Clear removes everything immediately.

Frequently asked questions

Does this tell me whether a name is available?: No, and it is important that it does not pretend to. It compares the names in your list with each other and nothing else. Whether a name is registered as a trademark, available as a domain, or already in use by somebody else are separate questions that need a search of the relevant register — and a similarity score is not evidence about any of them.
What does the score mean?: It is a rough 0-100 combination of five tests, sorted so the pairs most likely to be misread come first. Above 75 is labelled very close, above 55 close, and anything lower is worth a look. The number is there to order the list, not to be a measurement of anything in the world.
Why Soundex rather than something more modern?: Because it is defined precisely enough to be verified. Soundex has published reference values — Robert and Rupert are both R163, Ashcraft is A261 — so an implementation can be checked against something other than its author's opinion. More sophisticated phonetic algorithms exist, but a subtly wrong implementation of one is worse than a correct implementation of a crude one.
How many names can I check?: Up to 200, which is 19,900 pairs and still instant. The comparison is every pair against every other, so the work grows with the square of the list, but at these sizes that costs nothing.