🔤Word CounterReady

Count words, visible Unicode characters, sentences, paragraphs, and lines locally, then use an optional target, repetition context, and transparent pace estimates to plan a draft.

Writing pulse

Descriptive signals for revising, not a readability grade or a writing score.

Unique termscase-insensitive: 0
Average word lengthvisible characters: — chars
Average sentencedescriptive only: — words
Longest wordfirst if tied: —

Repeated terms

Case-insensitive groups; no language-specific stop-word list is removed.

Write or paste text to see repeated terms here.

Word target

Optional. Use it for a brief, assignment, draft, or script; the counter never treats a target as a quality score.

Target words

Silent reading pace

A planning estimate you can choose, not a measurement of comprehension or readability.

Speaking estimate

Uses 130 words per minute and rounds up to a whole minute. Rehearsal, pauses, pronunciation, slides, and interpretation can change a real delivery time.

How these counts work—and where they stop

Characters are visible Unicode grapheme clusters when the browser supports Intl.Segmenter, so a joined emoji or base letter plus accent is not blindly counted by UTF-16 string length.
Words and sentences use the browser's locale-aware boundary data. They are useful counts, not a language detector, grammar checker, or legal platform-limit check.
Paragraphs need a blank line between them; lines include empty and trailing physical lines. A site's own character limit may count UTF-16 units, bytes, or a different unit.
Word targets, repeated terms, and pace estimates support revision and planning; none determines writing quality.

Unicode notes that language-specific dictionary or locale tailoring can still affect word boundaries, especially for scripts without spaces. Read the Unicode Text Segmentation guidance.

Capability checkedVerified September 9, 2026

What works

Use Intl.Segmenter grapheme, word, and sentence boundaries when the browser supplies them, filtering word segments by isWordLike
Count joined emoji and combining marks as user-perceived grapheme clusters rather than UTF-16 code units
Group repeated terms case-insensitively after NFC normalization without applying an English-only stop-word list

Important limits

Word and sentence boundaries are language- and implementation-sensitive estimates; scripts without spaces can require dictionary or locale tailoring that this page does not choose for the visitor
A destination platform can count UTF-16 units, code points, bytes, tokens, or different word boundaries, so this result is not proof of a social-network, submission, SEO, or legal limit

Reviewed capability guide

What the Word Counter does—and does not do

Exact supported inputs, outputs, limits, and primary sources for the current implementation.

Accepted input

Up to 250,000 JavaScript string units of pasted or typed text in browser memory
An optional whole-word target from 1 to 1,000,000
A selected careful, standard, or fast silent-reading pace of 150, 200, or 250 words per minute

Produced output

Visible character (grapheme), non-whitespace character, word, sentence, paragraph, and physical-line counts
Unique/repeated-term context, average word and sentence length, longest word, optional target progress, and copyable summary statistics
Reading and speaking-time planning estimates plus an explicit warning when the browser lacks Unicode segmentation support

Operational limits

Text above 250,000 JavaScript string units is not analyzed, so a pasted document cannot freeze this local workbench
The silent-reading estimate is selectable only at 150, 200, or 250 words per minute; the speaking estimate uses 130 words per minute and rounds up
Paragraphs are blank-line-separated and physical lines include empty and trailing lines
No language detection, grammar checking, readability score, plagiarism analysis, document upload, or server-side persistence is provided

Known limitations

Word and sentence boundaries are language- and implementation-sensitive estimates; scripts without spaces can require dictionary or locale tailoring that this page does not choose for the visitor
A destination platform can count UTF-16 units, code points, bytes, tokens, or different word boundaries, so this result is not proof of a social-network, submission, SEO, or legal limit
Repeated terms are revision hints, not a readability, vocabulary, authorship, plagiarism, or writing-quality judgment
Browser-local processing avoids upload, but text can still remain in the visible editor, browser session, clipboard, or a downloaded copy until the visitor clears it

Standards and primary sources

Continue the workflow

Text Diff Checker →

Capability claims last reviewed September 9, 2026. Review dates describe the implementation audit, not content freshness by itself.

Reviewed workflow

Use the Word Counter with the right expectations

A text counter should not mistake a joined emoji for several characters or assume every language separates words with spaces. This workbench uses the browser’s Unicode-aware segmentation when available, labels its fallback when it is not, and distinguishes user-perceived characters from non-whitespace characters and word-like segments. It adds practical revision signals—without turning a word count, repeated term, or pace estimate into a score for writing quality.

When it is the right tool

Checking a draft with emoji, combining accents, multilingual text, or scripts that do not rely on spaces between every word
Tracking a brief, assignment, script, or article target without uploading the text to a word-count service
Planning an approximate read or spoken delivery time at an explicitly chosen pace
Spotting repeated word-like terms without silently discarding common English words from a multilingual draft

A deliberate choice

Tooleras keeps this page specific to the job above instead of claiming support for adjacent formats, platforms, or edge cases it cannot verify. Open the capability guide for the exact inputs, outputs, limits, and sources.

Checked examples

Visible characters, not UTF-16 length

é 👨‍👩‍👧‍👦 🇱🇹 你好 world

Counts grapheme clusters so a base letter plus accent, a joined family emoji, and a flag are not inflated by their underlying UTF-16 representation; words still use the browser’s word-boundary data.

A short talk with a target

A 780-word draft · 800-word target · standard 200 words/min pace

Shows the remaining words, rounds the read estimate up to a whole minute, and keeps the 130 words/min speaking estimate clearly separate from actual rehearsal time.

Need a different tool?

Browse the full catalogue — all free, with data handling documented per tool.

Browse all tools