Feature

Twelve error types,
each explained.

Spelling is the easy part. WriteX also catches grammar, morphology, semantics, style, punctuation, misspelled names, and violations of your own style guide, color coded in the text and counted in the sidebar.

Depth, not just breadth

Categories no Latin-language checker has.

Arabic is a root-and-pattern language, so WriteX checks morphology as its own category. It reads meaning, so semantics gets a category too. Every error type has its own color in the text and its own counter in the sidebar.

  • Hamza placement, case endings, number agreement, and the traps professionals still fall into.
  • English sentence correction: the whole sentence rewritten in one accept, with a visible diff.
  • Spanish coming soon, on the acquired Correcto models.
خمسة عشرة تفاحةخمس عشرة تفاحة
Grammar

Compound numbers 13 to 19: the first part disagrees with the counted noun in gender.

AcceptDismiss
The jerney of a thozand mels begin wiz on stebThe journey of a thousand miles begins with one step
Style

Sentence correction: five fixes applied in a single accept.

AcceptDismiss

Category by category

What each category actually checks.

Eight categories, and inside them the specific checks the engine runs. This is the full list.

Spelling

Spelling correction

Text free of spelling mistakes: both the rule-based errors and the ones your fingers make.

Standard spelling errors

Hamza forms, the soft alif, and taa marbuta versus taa mabsuta, so the rules are followed and the forms never get confused.

Typographical errors

Keyboard slips: a letter deleted, added, or transposed, plus characters that look or sound alike.

Grammar

Grammar correction

Each word's role analysed and the sentence structure reviewed, so the meaning you intended survives.

Sentence structure accuracy

Word order and placement evaluated across the sentence, with errors corrected in place.

Grammatical role of the word

Nominative, accusative, genitive, and jussive endings fixed from the sentence's meaning, word position, and the relationships between words.

Morphology

Morphology correction

Every word examined against its morphological pattern to separate root letters from the added ones.

Morphological structure

Errors from vowel and consonant letters, patterns with letters added or dropped, and mistakes in dual and plural forms.

Incorrect derivation

Words derived on the wrong pattern because a neighbouring form was easier or closer to hand.

Semantics

Language and semantics

Errors that depart from the usage eloquent speakers follow: context, meaning, vocabulary, and dialect mixed into standard Arabic.

Semantic and contextual errors

Accidental misreadings prevented: many everyday words carry meanings quite different from how they are commonly used.

Lexical errors

Gender, singular and plural, words used with the wrong sense, and the wrong preposition after a verb.

Dialectal errors

Informal language detected, with standard Arabic alternatives suggested to unify the writing.

Style

Style correction

Usage that drifts under outside influence, pulled back toward the most precise and flexible features of the language.

Stylistic errors

Weak constructions refined into more precise phrasing and more eloquent structures.

Clarity of structure

Synonyms surfaced for words a reader may find ambiguous or unfamiliar.

Linguistic politeness

Polite alternatives suggested so wording never embarrasses or harms the person it describes.

Punctuation

Formatting correction

The typesetting layer: punctuation, joined or split words, and the spaces between them.

Punctuation marks

Missing commas in long sentences, absent periods at the end, and the rest of the marks a print-ready text needs.

Word separation and joining

Words fused together by fast typing are separated, and words wrongly split are rejoined.

Spacing

Extra spaces between words removed, whether typing ran fast or slow.

Names

Names correction

Every kind of proper name recognized and verified: people, places, major entities, and scientific and literary terms.

Named entities

Prominent figures across politics, art, sport, culture, history, and religion standardized to one approved form.

Title errors

Titles of prominent figures corrected to the style official protocol requires.

Names of entities

Major organizations with global reach refined, in football, theatre, cinema, and the sciences.

Geographical names

Continents, countries, capitals, cities, landmarks, states, and provinces adjusted.

Style guide

Quotation correction

Famous texts referenced accurately, with quoted content matched against the original source.

Religious texts

Quranic verses, hadiths, and other religious texts examined word by word, because precise recall is hard and accuracy matters.

Legal texts

Articles inside constitutional and legal texts from most Arab countries reviewed.

Famous quotes

Information about quotations from constitutions, Quranic chapters and verses, and hadiths, to speed up writing and reading. Institutional accounts can add their own.

Three speeds

Fix one, fix a category, or Correct All.

Click a word to fix it. Click a category to clear every spelling error at once. Or press Correct All and watch the counters drain to zero. In Word, double-click any error card and the cursor jumps to the word.

CorrectionsCorrect all (10)
Spelling3
Grammar2
Morphology1
Semantics0
Style1
Punctuation2
Names1
Style guide0

Frequently asked questions

What are the 12 error types?

The engine distinguishes 12 error classes, surfaced in the editor as 8 color-coded categories: spelling, grammar, morphology, semantics, style, punctuation and formatting, proper names, and institutional style guide violations. Sub-checks include quotations, dates, and numbers.

How accurate is Arabic checking compared to Word?

On the company benchmark sentence, Microsoft Word flags 2 errors; WriteX flags 6, each with the grammatical rule explained. Arabic morphology and semantics are simply not modeled by generic checkers.

How does WriteX work?

It runs a chain of specialised AI and natural-language-processing components rather than one general model, which is what lets it classify an error, explain the rule, respond in real time, and run inside your own network.

Does it replace a human proofreader?

No. It is built to support proofreaders: it catches errors a reader can miss and removes most of the mechanical correction work, while editorial judgement stays with the person.

Every word your institution writes, checked.

Official letters to social posts, Arabic and English alike. One assistant that proofreads, enforces your style, and explains every correction, wherever your teams write.

12 error types · Institutional style guides · Works in Office, browsers, and the web editor