Spell Checker

Check English or Spanish spelling against a real dictionary

About Spell Checker

This spell checker handles English and Spanish text. It does not check Korean or Japanese.

It works in three layers. First, a short built-in list of common misspellings (about 40 for English, about 30 for Spanish). Second, a full Hunspell-format dictionary, downloaded to your browser the first time you check, that flags any word it doesn't recognize. Third, formatting rules: double spaces, a space before punctuation, and the same character repeated four or more times. Knowing which layer produced a warning tells you how much to trust it.

How the dictionary check works

Hunspell is the spell-checking engine used by LibreOffice, Firefox and many other programs. Its dictionaries come in two files: a .dic list of word stems and an .aff file of affix rules, so "walk" plus rules generates walked, walking and walks without listing each form. This tool loads those files with nspell, a JavaScript implementation of the Hunspell format, from a public CDN. If the download fails, only the built-in list and the formatting rules run.

A dictionary check answers exactly one question: is this sequence of letters a known word? It can't tell whether it is the right word. "Their going to the store" passes, because "their" is spelled correctly. "Could of", "loose" for "lose", "affect" for "effect" all pass. That is the gap between a spell checker and a grammar checker, which looks at context.

The opposite error is flagging correct words that aren't in the dictionary: surnames, brand names, technical terms, new slang. To cut down the noise, the tool skips words of one letter, words in all capitals up to five letters (treated as acronyms like NASA or HTML), and capitalized words whose lowercase form is known.

What "fix all" changes, and where it goes wrong

The corrected text applies the built-in misspelling list, the spacing fixes and the repeated-character fix. Words that are only unknown to the dictionary are highlighted with suggestions but never replaced automatically, because Hunspell's first suggestion is often wrong (for "hav" it can offer "hag").

The automatic rules are kept narrow to avoid false positives, but a few cases still need a look:

The repeated-character rule applies to letters only. Numbers such as "1000000" are left alone.
Deliberate repetition is still shortened: "Hmmmm" becomes "Hmm".
Replacements from the list keep the original capitalization, so "Tambien" at the start of a sentence becomes "También".
Spanish "echo" is not in the list, because it is a correct form of echar ("yo lo echo", I throw it). The flip side is that "echo" written by mistake for "hecho" is not caught.
A space before punctuation is removed only when it is a space or tab; a line break is kept. A space before a colon on the same line is still removed, which is wrong for French typography if you paste French text.

Spanish accents

The most common Spanish mistake is a missing written accent: tambien, despues, informacion, numero. Those unaccented forms aren't words, so the dictionary catches them reliably. Pairs where both forms exist are a different matter: si and sí, el and él, tu and tú, mas and más, and the verb forms like esta and está all depend on meaning, and a word-by-word checker cannot tell them apart.

Why Korean and Japanese are not supported

Hunspell-style checking assumes you can list the valid words of a language, perhaps with affix rules. Korean is agglutinative: particles and endings stack onto stems, so the number of valid word forms is effectively open-ended, and spacing rules are themselves a large part of Korean spelling. Japanese is written without spaces, so the text must first be segmented into words. Checkers for both languages usually rely on morphological analysis on a server. This tool runs entirely in the browser, so it leaves them out.

Worked examples

Misspellings and formatting

Input: I definately recieve it .
Output: I definitely receive it.

The two misspellings come from the built-in list; the double space and the space before the period come from the formatting rules. All four are applied automatically.

What slips through

Input: Their going to the libary tommorow.
Output: Their going to the library tomorrow.

"libary" and "tommorow" are both in the built-in list and fixed automatically. "Their" is a real word, so it passes; the their/they're mistake is invisible to a spell checker.

Deliberate repetition and numbers

Input: Hmmmm, the budget is 1000000 dollars.
Output: Hmm, the budget is 1000000 dollars.

Four or more of the same letter are shortened; digits are excluded from the repeated-character rule, so the number stays intact.

How to use it

Choose the language to check: English or Español.
Paste your text and run the check. The dictionary downloads on the first run.
Review each highlighted word and accept only the suggestions you agree with.

Frequently asked questions

Does it check grammar?

No. It checks spelling and spacing around punctuation only. Homophones like their/there and agreement errors are not detected.

A correct word is flagged. What should I do?

Ignore it. Proper nouns and technical terms are often missing from the dictionary, and unknown words are never changed automatically.

Does it accept British spellings like colour?

It depends on the spelling variant the loaded English dictionary uses. If a British form is flagged, it is highlighted but not changed.

Is it safe for text with numbers?

Yes. The repeated-character rule only looks at letters, so long numbers are not shortened.

Sources

Last reviewed: 2026-09-22

Privacy and how data is handled