open.text

Small, dependency-free text analysis utilities for browsers and Node.js.

open.text provides focused helpers for counting words, characters, sentences, and paragraphs, estimating reading time, and producing a compact summary of a piece of plain text.

Install

npm install @journey-to-code/open-text

Usage

import {
  countWords,
  countCharacters,
  countSentences,
  countParagraphs,
  estimateReadingTime,
  analyzeText
} from "@journey-to-code/open-text";

const text = "Hello world. This is open.text.";

countWords(text); // 5
countCharacters(text); // 31
countSentences(text); // 2
countParagraphs(text); // 1

estimateReadingTime(text);
// {
//   words: 5,
//   minutes: 0.022222222222222223,
//   roundedMinutes: 1,
//   wordsPerMinute: 225
// }

analyzeText(text);
// {
//   words: 5,
//   characters: 31,
//   charactersWithoutWhitespace: 27,
//   sentences: 2,
//   paragraphs: 1,
//   readingTime: {
//     words: 5,
//     minutes: 0.022222222222222223,
//     roundedMinutes: 1,
//     wordsPerMinute: 225
//   }
// }

API

countWords(text, options?)

Counts word-like segments in a string.

countWords("don't stop"); // 2

Options:

countWords(text, { locale: "en" });

The implementation uses Intl.Segmenter when available and a Unicode-aware fallback otherwise.

countCharacters(text, options?)

Counts user-perceived characters (grapheme clusters) when Intl.Segmenter is available, with a Unicode code-point fallback.

countCharacters("hello"); // 5
countCharacters("hello world", { whitespace: false }); // 10

Options:

• locale — locale hint used by Intl.Segmenter.
• whitespace — defaults to true; set to false to exclude whitespace.

countSentences(text, options?)

Counts sentence-like segments.

countSentences("Hello. How are you? Great!"); // 3

Sentence detection is intentionally heuristic. open.text is not a full NLP sentence tokenizer.

countParagraphs(text)

Counts non-empty paragraph blocks separated by one or more blank lines.

countParagraphs("One paragraph.\n\nAnother paragraph."); // 2

Both LF and CRLF input are normalized.

estimateReadingTime(text, options?)

Estimates reading time from word count.

estimateReadingTime(text, { wordsPerMinute: 200 });

Returns:

{
  words,
  minutes,
  roundedMinutes,
  wordsPerMinute
}

The default reading speed is 225 words per minute.

analyzeText(text, options?)

Returns a complete analysis object.

analyzeText(text, {
  locale: "en",
  wordsPerMinute: 225
});

Input behavior

All public functions require a string.

countWords(""); // 0
countWords("   "); // 0
countWords(null); // TypeError

The library does not silently coerce non-string input.

Scope

open.text analyzes plain text.

It deliberately does not parse or understand:

• Markdown
• HTML
• frontmatter
• source code
• URLs
• sentiment
• keywords
• readability grades
• LLM tokens

Those concerns belong in higher-level tools.

Runtime

• Node.js 18+
• modern browsers
• ES modules
• zero runtime dependencies

Development

npm test

License

MIT