How to count words and characters in Arabic and English text

Words are what Intl.Segmenter calls words. If the passage contains Arabic letters, the segmenter runs in Arabic. Spaces separate words. So do the ordinary punctuation marks, including the Arabic comma ، and the Arabic question mark ؟. “مرحبا بالعالم” is two words. “Hello world” is two words. A number or a URL counts as a word when the segmenter says it is one, which is usually what you want for a limit and occasionally not what a teacher meant by “a URL does not count”. The page does not have a secret second definition. The tile is the segmenter’s answer.

Characters are Unicode characters, which is the count a person means by “letters”. An Arabic letter is one. A Latin letter is one. An emoji is one. This is not the number of bytes on disk and not the number of UTF-16 units, both of which can make Arabic look twice as long as it is. Characters without spaces is the same count with whitespace removed, including spaces, tabs, and line breaks. Assignments that say “500 characters excluding spaces” want that tile. Assignments that say “500 characters” usually want the full character count, spaces included. Read the instruction before you pick the tile. The page shows both so you do not have to guess and then recount by hand.

Sentences use the sentence segmenter. “مرحبا بالعالم. كيف حالك؟” is two sentences. A block of text with no terminator is one sentence, which matches the way a single line without a full stop is still one utterance. Paragraphs are separated by a blank line. A single line break inside a wrapped paragraph does not create a new paragraph. If you pasted from a PDF and every line is its own paragraph, the paragraph count will look high. That is the paste, not a bug. Join the lines if you wanted one paragraph.

English is estimated at 200 words a minute. Arabic is estimated at 180. Those are planning numbers for silent reading of ordinary prose, not a test score and not the speed of reading aloud. Mixed text uses a weighted average: each Arabic word contributes at 180 and each other word at 200. Under one minute, the label says “Less than a minute” instead of rounding up to 1 and making a sentence look like a page. At a minute and above, it rounds up, so 201 English words is two minutes. Use it to see whether a caption is a glance or a paragraph. Do not use it to promise a client that a reader will finish in that many minutes.

The counter does not know the subject. A legal clause and a chat message with the same word count do not take the same attention. The number is a size, not a difficulty.

Paste the passage into Word Counter. The box follows the text direction, so Arabic sits on the right and English on the left, and a mixed line follows the characters in it. Read the tile that matches the limit you were given. If you are drafting toward a limit, leave the page open and edit in the box, or edit in your document and paste again. The page does not sync with a file. It measures whatever is in the box.

The box stops counting after 100,000 characters and says so. That is a full chapter for most purposes. Count a longer work in parts and add the totals. The cap keeps the tab responsive. It is not an upload limit, because there is no upload.

Tashkeel, the short-vowel marks, are characters. They are included in the character counts. They do not each become a word. If your limit is on words, vowel marks will not inflate the word tile. If your limit is on characters and the passage is fully vocalised, the character tile will be higher than an unvocalised copy of the same words. Count the version you will submit.

If the task was not to count an amount but to spell it — “one thousand five hundred riyals” in Arabic for a contract — that is a different tool. Arabic Numbers writes the words with currency grammar. The counter would only tell you how many words that line contains after you have it.

It will not tell you whether a hyphenated word is one word or two beyond what the segmenter does. It will not ignore quotations, headings, or footnotes unless you delete them from the box first. It will not detect translation direction or dialect. Those are editorial choices. Make them before you paste, or delete the parts that do not count, and then trust the tiles for what remains.

It also will not store the text. That is worth remembering when the passage is a draft you would not post. The counting runs as you type. Closing the tab clears it. The public site can still record that the page was opened, which is described in the Privacy Policy and in privacy with online file tools. Do not paste passwords or full identity numbers into a word counter. You do not need their word count.

Why is my Arabic character count lower here than in another tool?

The other tool may be counting UTF-16 units, two per Arabic letter. This page counts characters. For a limit set by a person, characters are the right unit. For a limit set by a database column, ask which unit the column uses before you rely on either number.

Does a line break count as a character?

Yes, in the full character count. It is excluded from “without spaces”. It does not by itself start a new paragraph unless there is a blank line.

Can I count only the Arabic and ignore the English?

Not as a separate mode. The tiles include everything in the box. The line under them tells you how many of the words were Arabic and how many were not, which is enough to see the mix. Delete one language from the box if a limit applies to only one of them.

Is the reading time in minutes of speech?

No. It is a silent-reading estimate. Reading aloud is slower, and a presentation is slower again.

Open Word Counter and paste the passage you are actually submitting.