1. Home
  2. Text tools

Word and character counter

Exact counts as you type. Your text stays on your device.

0Words
0Characters
0Without spaces
0Sentences
0Paragraphs
0 sReading time
0 sSpeaking time
0Lines
0Bytes (UTF-8)
Length limits
X (Twitter) post · 280280 left
SEO title (approx.) · 6060 left
Meta description (approx.) · 155155 left
SMS (GSM-7, one part) · 160160 left
Instagram caption · 2,2002,200 left
LinkedIn post · 3,0003,000 left

Your text never leaves this device.

Paste or type your text to see the number of words, characters with and without spaces, sentences, paragraphs and lines instantly, along with reading and speaking time. Counting follows the rules of each language, so Chinese, Japanese, Arabic and Turkish are counted correctly. The text is never sent anywhere.

How to use

  1. Type or paste your text.
  2. Read the counts as they update.
  3. Pick a limit, such as an X post or a meta description, to see what is left.

Frequently asked questions

How are words counted?

Words are found with the browser's built-in language rules (Unicode word segmentation), not just spaces. Numbers count as words, punctuation does not, and languages without spaces such as Chinese and Japanese are split into real words.

How do I count characters without spaces?

Look at “Characters without spaces”. Spaces, tabs and line breaks are left out, while letters, digits, punctuation and emoji are counted.

How many words fit on one page?

Roughly 500 words single-spaced or 250 double-spaced with a 12 pt font, but the real number depends on font, margins and spacing, so treat it as an estimate.

How is reading time calculated?

Reading time assumes 238 words per minute, the average silent reading speed for adults found in a large 2019 review of studies. Speaking time assumes 150 words per minute.

What is the character limit of an X post or a meta description?

A standard X (Twitter) post allows 280 characters. Search engines usually show about 150–160 characters of a meta description, so keeping it under 155 is a safe target.

Is word counting different in Turkish, Chinese or Arabic?

Yes. Turkish suffixes after an apostrophe stay in one word, Arabic is counted by its own rules, and Chinese and Japanese are split with a dictionary because they have no spaces. This counter handles each correctly.

Why do AI chatbots get word counts wrong?

Language models read text as tokens, which are pieces of words, so they estimate counts instead of counting. For exact numbers use this counter, or connect it to your assistant through MCP.

Should I count words or characters?

It depends on who sets the limit. Essays, articles and briefs use words. Social posts, ad copy and form fields use characters. The page has both, plus a focused page for counting characters. Look at "characters without spaces" only when a rule says so; otherwise the normal count includes spaces.

The built-in limits show how much you have left: 280 for an X post, about 60 for an SEO title, about 155 for a meta description, 160 for an SMS, 2,200 for an Instagram caption and 3,000 for a LinkedIn post. An emoji counts as one character.

How do I check my text for search engines?

Write the title and the description separately and paste each in turn. Aim for about 60 characters for the title and under 155 for the description, so search engines are less likely to cut them off. These numbers are guides, not hard rules, because the visible width also depends on the letters. For a headline with a nice rhythm, change its case and check the count again.

How are words and pages worked out?

Words are found with the browser's language rules, not by splitting on spaces. That is why Chinese, Japanese, Arabic and Turkish text with suffixes after an apostrophe are counted sensibly. Reading time assumes 238 words per minute and speaking time 150. A page estimate of about 500 words single-spaced is rough, since fonts and margins change it.

What mistakes should I avoid?

Check which count the requirement means. A limit of 200 characters can include or exclude spaces, and the two numbers differ a lot. If a database field has a size limit in bytes, read the UTF-8 byte count, because accented letters take more than one byte. When you edit a draft and need to see what changed, put both versions into Text Compare instead of counting by eye. And if a chatbot gave you a count, check it here: language models estimate instead of counting.

How it compares

This toolUpload-based services
Is your text sent to a server? No Usually
Account required No Sometimes
Length or usage limit None Often
Saved or logged Never Possibly

Limits and privacy

  • Page estimates are approximate because they depend on fonts and layout.
  • Reading and speaking times are averages; real speed varies by person and text.
  • Very long texts (several megabytes) may update more slowly.

Standards and sources

Use it in your AI assistant

Language models count tokens, not words, so their counts are guesses. The count_text MCP tool, available remotely and locally, gives an assistant exact numbers, for example to keep a summary under 150 words or a post under 280 characters.

Set up the MCP server

Popular uses

Updated: Something not working? Report a problem