阅读中文版本

HTML to Plain Text CLI

A dependency-free Node CLI that converts a small HTML document to readable plain text, honoring block breaks and decoding common entities.

node --version

node html-to-text.mjs page.html

node --test html-to-text.test.mjs

Prints the visible text with block elements on separate lines and HTML entities decoded, and strips script and style contents.

What it does

Removes script and style blocks, replaces block-level tags with newlines, strips remaining tags, and decodes a set of named and numeric entities. It is intentionally a small, predictable converter rather than a full parser.

const ENTITIES = { amp: '&', lt: '<', gt: '>', quot: '"', apos: "'", nbsp: ' ' };
const BLOCK = /<\/?(p|div|h[1-6]|li|tr|br|section|article)[^>]*>/gi;

export function htmlToText(html) {
  return html
    .replace(/<(script|style)[\s\S]*?<\/\1>/gi, ' ')
    .replace(BLOCK, '\n')
    .replace(/<[^>]+>/g, '')
    .replace(/&#(\d+);/g, (_, d) => String.fromCodePoint(Number(d)))
    .replace(/&(\w+);/g, (m, name) => ENTITIES[name] ?? m)
    .split('\n').map((l) => l.trim()).filter(Boolean).join('\n');
}

import { readFileSync } from 'node:fs';
const file = process.argv[2];
const input = file ? readFileSync(file, 'utf8') : readFileSync(0, 'utf8');
console.log(htmlToText(input));

Run and verify

node html-to-text.mjs page.html

Run the test command to check the implementation: node --test html-to-text.test.mjs.

Expected Output

For a document containing <h1>Title</h1><p>A &amp; B</p>, script and style are dropped and entities decoded:

Title
A & B

Limitations

Not a spec-compliant parser: malformed markup, entity edge cases, and nested tables may not convert as a browser would render them.