Count words in text, as you write
- Type, paste, or import a UTF-8 TXT or Markdown file. Your word and character counts update as the text changes.
- Set a word or character target, or open Advanced writing statistics for phrase frequency, timing settings and an optional English readability estimate.
- Copy or download your text before leaving. Browser autosave is optional and starts off.
What counts as a word?
This word count checker uses Intl.Segmenter with English locale rules and Unicode word boundaries. Letters and numbers count; punctuation and emoji alone do not. Internal apostrophes stay inside words, so “don’t” is one word. Hyphens (-, ‐ and ‑) split words: “well-known” counts as two. The same rules apply to the overall count and frequency analysis. Browser Unicode dictionaries may differ for some languages and abbreviations.
Unicode, characters and whitespace
Characters are Unicode code points, including spaces, tabs and line breaks. “Without whitespace” removes Unicode White_Space characters and the byte-order-mark character first; zero-width joiners and zero-width spaces still count. An emoji can contain several code points, and a letter with a combining accent can contain two, even when it looks like one symbol. Windows CRLF and old Mac CR line endings become a single LF. The editor limit is 100,000 UTF-16 units, a separate browser length measure; emoji can use two units per code point.
Sentences, paragraphs and lines
Sentence boundaries use Intl.Segmenter within each physical line. Each segment containing a word counts as a sentence, including an unfinished sentence. Each nonempty line therefore has at least one sentence; punctuation-only lines have none. Blank lines separate paragraphs. Lines count every physical line, including a trailing empty line; an empty editor has zero. These are text statistics, not a grammar check.
How reading and speaking time are estimated
Reading time starts at 200 words per minute and speaking time at 130. We divide the word count by your chosen rate, multiply by 60 and round up to a whole second. Change both speeds in the advanced panel. Pauses, familiarity, language and presentation style affect real timing. Estimated pages divide words by your editable words-per-page setting, initially 250; no page formatting is measured.
An essay word counter with editable targets
Paste only the parts your assignment includes. This tool counts everything in the editor, including headings, citations and references; it cannot infer your assignment’s rules. Set any whole-number word or character target from 1 to 1,000,000. Remaining counts stop at zero and over-target warnings show the excess. Character goals include whitespace. Presets are editable starting points, not official assignment or platform limits.
Understand keyword frequency, without an SEO score
Frequency combines case variants, NFC-equivalent spelling and straight or curly apostrophes. One-word density is occurrences divided by all words. Two- and three-word density uses all windows of that length within each sentence or line; overlapping windows count. Excluding common English words hides any phrase containing one, without joining across the removed word or changing totals or density denominators. The top 15 results are shown. There is no universal ideal keyword density or ranking guarantee.
Readability is a rough English estimate
Opt in only for English text with at least 100 words and three sentences. The Flesch Reading Ease formula is 206.835 − 1.015 × words per sentence − 84.6 × syllables per word. This tool estimates syllables with vowel groups and a small exception list. Names, abbreviations and unusual spellings can be inaccurate. Non-basic-Latin word forms or numeric tokens make the estimate unavailable. A higher score generally suggests easier reading, not better writing; scores can fall outside 0–100.
Unique words and sentence lengths
Unique words combine case variants, NFC-equivalent spelling and straight or curly apostrophes. Average word length counts letter, combining-mark and number code points within tokens, excluding punctuation such as apostrophes. Average sentence length divides words by counted sentences. Longest and shortest sentences are chosen by word count, not characters; tied results show one matching sentence. Previews stop at 200 UTF-16 units, while statistics use the full text.
Browser privacy and safe text editing
Analysis and file processing happen on your device, without an account, upload or text analytics. Optional autosave stores your text and settings in this browser profile, accessible to anyone using it. Delete saved copy removes that data without emptying the editor. Clear text asks first and also disables autosave. Downloads are recommended for important writing. Case changes, imports and clearing retain up to ten prior versions for Restore previous text during this visit; refresh ends that history. Normal typing uses your browser’s undo and redo.
Formatting cleanup that you control
Case tools change the selected style for all text, including names and acronyms. Review the result and restore it if needed. Remove extra spaces collapses repeated ordinary spaces between non-whitespace characters, preserving indentation, tabs and trailing Markdown spaces. Remove extra blank lines reduces runs to one blank line. Neither tool rewrites your sentences. Markdown is counted as plain text, never executed or rendered.
Try these worked examples
Words versus characters
“Hello world!” has 2 words, 12 characters including the space and 11 without whitespace. At the default speeds, reading and speaking time both round up to 1 second. A 10-word target has 8 words remaining and 20% progress.
Apostrophes, hyphens and emoji
“Don’t stop. Well-known words 😊” has 5 words and 2 sentences. The emoji contributes 1 code point and no words. A family emoji joined with invisible joiners contributes several code points, even though it looks like a single picture.
Overlapping phrase frequency
“red fox red fox” has 4 words. “red fox” occurs twice among 3 two-word windows, so its two-word density is 66.67%. As single words, “red” and “fox” each occur twice and each has 50% density.
Timing a draft
A 500-word draft takes an estimated 2 min 30 sec to read at 200 words per minute, or 3 min 51 sec to speak at 130. At 250 words per page, its estimated length is 2 pages.
Word counter questions
Is this online word counter free?
Yes. Counting, goals, frequency analysis, text utilities and TXT downloads are available without registration or a paid API.
Will counts match Microsoft Word or Google Docs?
Not necessarily. Applications use different word, punctuation and document rules. Taskono does not claim identical counts. Use the counter required by your assignment or publisher when exact agreement matters.
Can it count Chinese, Japanese or other non-Latin text?
Modern browsers use Unicode segmentation, including dictionary boundaries where supported. In browsers without Intl.Segmenter, the fallback counts letter/number runs with internal apostrophes, and may undercount languages without spaces. Its sentence fallback splits on common English and CJK sentence punctuation and line breaks. A visible notice identifies fallback mode. English stop-word filtering and readability are not multilingual analysis.
Can I import a PDF or Word document?
This version supports UTF-8 .txt and .md files only, up to 1 MB and the editor length limit. PDF, DOCX and legacy encodings are not supported. Copy plain text from your document or save it as UTF-8 text first.
What if browser storage is blocked or full?
The editor keeps working and shows a saving warning. Your current text remains in the tab, but it may not survive refresh. Download or copy it before leaving. Unreadable saved data is left untouched unless you agree to replace it or choose Delete saved copy.
Does the tool send my writing anywhere?
Taskono does not upload editor text or imported files, and has no text analytics. Optional saving is browser-local. Hosting still receives normal page requests; your browser extensions and device settings are outside Taskono’s control. Avoid browser autosave on a shared profile.