electron-spellchecker

Modified version of electron-spellchecker that is used in Mark Text.

electron-spellchecker is a library to help you implement spellchecking in your Electron applications. This library intends to solve the problem of spellchecking in a production-ready, international-friendly way.

electron-spellchecker:

• Spell checks in all of the languages that Chromium supports by reusing its dictionaries.
• Support for native macOS and Windows 10 spell checker that are prefered by default.
• Automatically detects the language the user is typing in and silently switches on the fly.
• Handles locale correctly and automatically (i.e. users who are from Australia should not be corrected for 'colour', but US English speakers should)
• Checks very quickly, doesn't introduce input lag which is extremely noticable
• Only loads one Dictionary at a time which saves a significant amount of memory
• Passive mode: don't underline spelling mistakes but you are still able to test for spelling mistakes.
• Custom user dictionary for both Hunspell and macOS.
• Option to ignore a word at runtime only.
• Attach spellchecker to a custom DOM element instead document.body.

Features removed:

• Dropped support for <=Electron 4.
• Removed context menus.
• Dropped automatic dictionary download and management.

Quick Start

import {SpellCheckHandler} from 'electron-spellchecker';

const cacheDir = '';

window.spellCheckHandler = new SpellCheckHandler(cacheDir);
window.spellCheckHandler.attachToInput();

// Start off as US English, America #1 (lol)
window.spellCheckHandler.switchLanguage('en-US');

Language Auto-Detection

The spell checker will attempt to automatically check the language that the user is typing in and switch on-the fly. However, giving it an explicit hint by calling switchLanguage, or providing it a block of sample text via provideHintText will result in much better results.

Sample text should be text that is reasonably likely to be in the same language as the user typing - for example, in an Email reply box, the original Email text would be a great sample, or in the case of Slack, the existing channel messages are used as the sample text.

About node-spellchecker

This module uses a fork of Atom's excellent node-spellchecker with few modifications. You can find the source here.