A word counter that counts the way you read
Type or paste text and every number updates instantly: words, characters with and without spaces, sentences, paragraphs, lines, unique words, reading time and speaking time. Below the editor you get live meters for the limits that matter when you publish — an X post, an Instagram caption and its 30-hashtag cap, a YouTube title and description, an SEO meta title and meta description, a LinkedIn post, a Pinterest description and an SMS — plus a keyword-density table for one-, two- and three-word phrases.
The counter runs entirely in your browser. Your essay, script, client copy or unpublished article is never uploaded, and the draft you are typing is saved only in this browser so a refresh does not lose it.
Correct counting for Hindi and other Indian languages
Most counters measure “characters” as JavaScript string length, which counts every invisible Unicode piece. In Devanagari, a single visible letter such as क्षि is made of several code points (क + ् + ष + ि). A typical tool reports four characters; a reader sees one syllable. This tool counts grapheme clusters — the letters a person sees — using the Unicode segmentation rules built into modern browsers, so Hindi, Marathi, Nepali, Sanskrit, Bengali, Tamil and Punjabi text get sensible character counts. For technical needs you also see Unicode code points and UTF-8 bytes. Sentences ending in the danda । and double danda ॥ are counted correctly, as are Japanese and Chinese sentences ending in 。.
How each number is calculated
- Words: runs of letters, marks and digits, with apostrophes and hyphens kept inside a word, so don’t and state-of-the-art count as one word each. Japanese, Chinese and Thai, which do not use spaces, are split with the browser's dictionary-based word segmenter.
- Characters: visible characters including spaces and line breaks; characters (no spaces) removes all whitespace.
- Sentences: text ending with . ! ? । ॥ … or their full-width forms. Common abbreviations (Mr., Dr., e.g., i.e., Rs.) and decimals such as 3.5 do not end a sentence.
- Paragraphs: blocks of text separated by at least one blank line.
- Reading time: words ÷ reading speed. The default 238 words per minute comes from a 2019 meta-analysis of 190 studies by Marc Brysbaert of silent reading of English non-fiction; choose 200 for technical or second-language reading and 300 for light reading.
- Speaking time: words ÷ 130 wpm, a comfortable presentation pace. Pick 110 for slow narration or 160 for a fast podcast.
- Flesch reading ease and grade level: standard formulas based on words per sentence and syllables per word. They are only meaningful for English, so they appear only for English text of 30 words or more.
Social media and SEO limits
X counts posts with weights: most Latin, Devanagari and other scripts count 1 per character, links always count 23, and emoji or Chinese, Japanese and Korean characters count 2. Our meter applies the same rules, so the number matches what the X composer shows. Instagram allows 2,200 characters and 30 hashtags in a caption. YouTube allows 100 characters in a title and 5,000 in a description; only the first ~70 title characters show in most search results, so put the key phrase first. For SEO, aim for a meta title of about 50–60 characters and a meta description of about 150–160 characters. Google actually truncates by pixel width, so treat these as guides rather than hard rules.
SMS length and why Hindi messages cost more
An SMS using only the GSM-7 alphabet holds 160 characters, or 153 per part when the message is split. As soon as it contains any character outside that alphabet — Hindi, an emoji, or even curly quotes — the whole message switches to Unicode and the limit drops to 70 characters (67 per part). The SMS meter shows which encoding applies and how many parts will be billed, which is useful for OTP templates, DLT-registered promotional messages and school or clinic reminders.
Using keyword density without keyword stuffing
The density table lists the most frequent words and phrases with their share of all words. Common words like the, and, के and है are skipped by default; untick the option to include them. There is no magic percentage that Google rewards. Use the table to check that your main topic is actually present, to spot accidental repetition, and to find synonyms you could use instead. If one phrase is far above everything else, read the text aloud — it probably sounds repetitive.
Tips for students, writers and creators
- Set a word goal for an essay, assignment or blog post and watch the progress bar.
- Use speaking time to time a speech, a YouTube script or a viva presentation before you record.
- Paste subtitles (.srt or .vtt) to estimate how long a video's dialogue runs.
- Press Ctrl+Shift+C to copy a summary of the stats into a report or a client email.
Frequently asked questions
Does the counter save or upload my text?
It never uploads. The current draft is stored only in your browser's localStorage so a refresh keeps it; Clear moves it to the private history, which you can delete.
Why is my Hindi character count lower than in other tools?
Other tools count Unicode code points, so a conjunct like क्षि counts as four. We count visible letters (grapheme clusters). The code-point count is shown under More statistics if you need it.
How is reading time calculated?
Word count divided by 238 words per minute, the average silent reading speed for English non-fiction found in a 2019 meta-analysis. You can change the speed.
Does the X meter match the app?
Yes, it uses X's weighted rules: links count 23, emoji and CJK characters count 2, and most other characters count 1.
How many characters can an SMS have in Hindi?
70 characters in one message, or 67 per part when it is split, because any non-GSM character switches the message to Unicode.
What keyword density is best for SEO?
There is no fixed ideal. Make sure the topic appears naturally, avoid repeating one phrase much more than others, and write for readers first.