How many homophone pairs does the detector know?

Forty-four documented pairs, each with its own context rule. They include your and you're, its and it's, their, there and they're, then and than, to, too and two, whose and who's, affect and effect, weather and whether, and lose and loose.

Why does it flag your going but not your book?

Because the rule is written around the construction, not the word. A possessive followed by a gerund is almost always meant to be you are, while your book is exactly right. Only the first pattern exists in the rule list.

Will it rewrite my text?

No. Every finding shows the word as written and the spelling the context calls for, side by side. Nothing changes in the draft unless you change it, so a finding you disagree with costs you nothing.

Why are some homophones not checked at all?

Because no context rule would be reliable. Where both spellings fit the same surroundings, a rule would fire on correct sentences, and a checker that cries wolf gets ignored. Those pairs are excluded rather than guessed at.

Can it tell that a flagged word is deliberate?

No, and it does not claim to. Dialogue, puns, idioms and quoted material can all put a flagged form where a rule does not expect it. Read each finding against the sentence and overrule it whenever the text is right.

Homophone Confusion Detector

Your text

Paste or type text

Result

Start typing above — the results update as you write.

What is checked, and how

44 rules cover 44 documented pairs: your / you're, its / it's, their / there / they're, then / than, to / too / two, who's / whose, affect / effect, weather / whether, lose / loose, and their relatives. Each rule is a construction, not a word: the tool fires on "your going" but stays silent on "your book".

A finding is a question, not a correction. The tool never edits your text — it shows the word, the spelling the surrounding words call for, and why. Read the sentence and overrule it whenever the flagged form is what you meant.

Pairs whose contexts are identical are not checked at all. "I read it yesterday" and "I red it yesterday" cannot be told apart by the words around them, so a read / red rule would only produce noise — the list leaves those out.

The Homophone Confusion Detector checks 44 documented pairs — your and you're, its and it's, their and there, then and than — but only where the surrounding words make one spelling implausible, which is why it can name a reason instead of just a count.

What is Homophone Confusion Detector?

Homophones sound alike and are spelled differently, which is exactly the kind of mistake a spelling checker cannot see: your and you're are both real words, so both pass. This homophone confusion detector looks at the words around each candidate instead of the candidate alone, and it publishes the pairs it knows so the claim is checkable rather than mystical. It is a narrow answer to a narrow problem: word confusion between two real words, which no dictionary can resolve. There are 44 pairs and 44 rules, one per pair.

Every rule is a construction. your going is flagged because a possessive is followed by a gerund; your book is not, because that is correct. more then enough is flagged because than is the comparison word and then is the time word, while and then we left stays clean. has past is flagged, has passed is not. Each finding carries the word as written, the spelling the context calls for, the pair it belongs to, and a one-line reason, so you can agree or overrule in a second.

The alternative — replacing every occurrence of than with then or vice versa — would be worse than useless. A blanket rule turns correct sentences into new errors, and the tool never does it: it reports, it never edits. That is also why the pair list is selective. Where two homophones fit the same context perfectly, no rule is written at all. I read it yesterday and I red it yesterday cannot be separated by the words around them, so no amount of rule-writing would help, and a noisy rule would train you to ignore the useful ones.

What the detector cannot see is worth stating plainly. It has no idea what you meant; it knows constructions, not sentences. A deliberate pun, a quoted line of dialogue, an idiom or a regional usage can all put a flagged form in a place the rule does not expect, and a contraction inside a name or a title may be missed. The findings are questions to answer, never corrections to accept blindly.

In practice the tool earns its place because these errors survive every other check. A spell check passes them, a grammar check often ignores them, and the eye slides over them after the third reading, because the wrong form is a real word and looks plausible. Seeing them listed with the reason attached is usually enough to fix a habit permanently — most writers find one or two pairs account for nearly everything. A finding that repeats in the same construction is usually a phrase you type without reading, and rewriting that phrase once removes the whole family of hits.

Run it before you run a spell check, or after, but run it on the final text rather than a first draft, because a sentence you are still restructuring may lose the construction that was flagged. Copy report gives you the list with the reasons, so the decisions can be made away from the screen, and the pair summary at the end shows which pairs actually mattered rather than leaving you with a total to interpret. Nothing is uploaded, and the text disappears with the tab.

How to use Homophone Confusion Detector

Paste the draft; the 44 pair rules run across the whole text as you type.
Read each finding with its reason: the word found, the pair, and the spelling the surrounding words call for.
Check the sentence in context before changing anything — the tool reports, it never edits.
Keep the findings you disagree with and note the pair; repeated habits show up clearly in the pair summary.
Copy the report and re-run after editing, since rewriting a clause can remove the construction that fired.

When to use Homophone Confusion Detector vs related tools

Use this detector on any draft where a wrong real word would embarrass you, and pair it with the checks that cover different word-level faults. The Apostrophe Error Finder handles dropped and stray marks in contractions and possessives, including the forms this tool leaves alone on purpose, and the Spell Checker covers words that are not in the dictionary at all.

When the confusion is a matter of sentence structure rather than a pair, the Grammar Checker runs a published list of mechanical checks over the same text.

Privacy & Security

This tool runs entirely in your browser — no data ever leaves your device. There is no server round-trip, no upload, no logging, and no account required. Your input is processed locally using client-side JavaScript and is never stored, transmitted, or accessible to anyone else. When you close the tab, everything disappears.