Word Counter & Character Count

What each box counts, and why two counters disagree

Every number above updates on each keystroke. Counters disagree with each other because "length" is not one measurement: a word processor, a tweet composer, a database column and a grader with a word limit each count something different. The first job is knowing which figure the person asking wants.

• Characters counts Unicode code points: every letter, digit, punctuation mark, space and line break, once each. This is what nearly every social platform and form field means.
• Without spaces is the same text with whitespace stripped, including the non-breaking space and the ideographic space. It does not strip the zero-width space, which Unicode does not class as whitespace, so text padded with invisible characters counts higher than it looks.
• Wordscounts runs of letters in any script that puts spaces between words: Latin, Cyrillic, Greek, Arabic, Hebrew, Devanagari, Korean hangul and the rest. A hyphenated compound such as state-of-the-art is one word, and so is don’t, with a straight or a curly apostrophe. Accented letters stay inside their word, so café and naïve count as one each. A standalone number such as 2026, 3.14 or 1,000 is one word too, as it is in Microsoft Word and Google Docs. Chinese, Japanese, Thai, Lao, Khmer and Burmese are written without spaces, so their text adds nothing to Words.
• Han characters counts Han characters only, the script shared by Chinese, Japanese kanji and Korean hanja, excluding punctuation. Japanese kana and Korean hangul are not Han, so they are not in this box. Length in Chinese is quoted in characters, not words.
• Sentences, paragraphs and lines split on sentence punctuation, blank lines and line breaks. An abbreviation such as e.g. reads as a sentence ending, so treat Sentences as a structural check, not an exact figure.
• UTF-8 bytes is what storage costs: one byte per ASCII letter, two for most accented Latin and Greek, three for most CJK characters, four for most emoji.

Emoji are where the arithmetic surprises people. A thumbs-up is one code point and counts as one. A national flag is two regional indicator letters and counts as two. A family emoji is a sequence joined by invisible characters and can count as five or more. You see one picture; the counter, the database and the platform limit see the sequence.

Which number you actually need

• A post against a platform limit: Characters, but read the platform section below, because X does not count the way this box does.
• An essay, a personal statement, a report: Words, unless the brief says characters.
• A brief that says "excluding spaces": Without spaces. Ask whether footnotes, captions and references are inside the count; that differs by department.
• A database column, an API field, a filename: UTF-8 bytes. A field advertised as 255 characters is usually 255 bytes.
• A page title or meta description: Characters, as a rough guide only. Search engines cut by rendered width, not by character count.
• Reading or speaking time: Words, divided by 200 to 250 for silent reading.
• Chinese text: Han characters. The Words box counts only the Latin words and numbers mixed into it.
• Japanese text: Characters or Without spaces. Han characters counts the kanji only and skips hiragana and katakana.
• Korean text: Words for a word limit, Characters or Without spaces for a character limit. Hangul is not counted in the Han characters box.
• Checking the shape of a draft: Paragraphs and Sentences. Twelve sentences in one paragraph is a wall of text at any word count.

How to check a post before you publish it

1. Write or paste the post into the box above. Include the line breaks you plan to keep: they are characters too, and a caption laid out with blank lines can cost twenty characters you did not budget for.
2. Read the Characters box and compare it with the limit for that platform in the list below.
3. If you are close to the limit, cut the parts that cost most per unit of meaning: emoji that repeat the words, trailing hashtags nobody follows, and the phrase "in this post I want to talk about".
4. Paste the final text into the real composer and trust its counter for the last few characters. It knows its own rules, including how it rewrites links.

How to hit an essay or application word count

1. Paste only the text that counts. If the brief excludes the title page, the abstract, footnotes or the bibliography, leave them out rather than counting the whole document and guessing the difference.
2. Read Words. Many institutions allow a margin of around ten percent either side, but plenty enforce a hard ceiling and simply stop reading at the limit, so confirm the rule in your own brief before relying on a margin.
3. Over the limit, cut whole sentences rather than trimming adverbs. The Sentences and Paragraphs boxes tell you whether you removed structure or just texture.
4. Under the limit, add evidence, not words. Padding is visible at exactly the length a grader reads every day.

Common targets, current at the time of writing: the Common Application personal essay allows 650 words with a 250 word minimum; the UCAS personal statement is capped at 4,000 characters including spaces, though the questions that allowance is split across changed recently, so check the rules for your own cycle; journal abstracts are commonly capped somewhere between 150 and 300 words. For timing a talk, 200 to 250 words a minute is the usual estimate for silent reading and a spoken delivery is slower, so a ten minute slot is closer to 1,300 words than to 2,500.

How each platform counts, at the time of writing

• X: 280 weighted characters, not 280 raw ones. Latin letters, digits and common punctuation weigh one; CJK characters, most other scripts and emoji weigh two. An emoji assembled from several code points still weighs two rather than one per code point. Every link is rewritten to a fixed-length t.co address, so a long URL costs no more than a short one. The developer documentation sets out the weighting in full.
• Instagram: captions up to 2,200 characters, bio up to 150. Only the opening of a caption, somewhere around 125 characters, shows in the feed before the "more" link, so the first line does the work.
• LinkedIn: posts up to 3,000 characters, with the opening lines collapsed behind "see more".
• YouTube: titles up to 100 characters, though the visible portion in search and on mobile is far shorter.
• Google search results: aim for about 60 characters of title and about 155 of meta description. Both are conventions rather than limits. Google truncates by pixel width, so a title in wide capitals is cut sooner than one in narrow lowercase, and it may rewrite the title entirely.
• Forms and databases: a field described as "max 255 characters" is usually 255 bytes. Accented Latin, CJK and emoji each cost more than one, which is why a form can reject text that looks well inside the limit. Use the UTF-8 bytes box for these.

Limits worth knowing about

This is not the counter that judges you. When a submission portal or an instructor opening your file in Word enforces the limit, that tool decides. Draft here, confirm there.

Invisible characters inflate the count. Zero-width characters pasted in from elsewhere are counted, are not removed by Without spaces, and look like nothing at all. If Characters is higher than you can explain, that is usually the reason. The invisible text page explains where they come from and how to strip them.

Screen readers. The counters are plain text beside their labels rather than a live region, so they are not announced automatically as you type. Move to them when you want the current figure.