💡Fresh releases: Style guide, updated language engine and custom dictionaries

Dear WebSpellChecker clients!

New season, new features, new perks. This blog post introduces you to updates and enhancements we at WebSpellChecker have been on lately.

So, without further ado, let’s see what we’ve found at the end of the WebSpellChecker rainbow this spring.

Style guide with custom rules

Language matters.

We were marching through last months with an idea of how our digital solution could contribute to effective business communication and corporate culture. As a result, we’ve developed and implemented the functionality of Style guide with custom rules. It allows users to add custom linguistic rules and patterns for smooth internal and external interactions.

In the latest product release, we introduced custom rules for non-inclusive and profanity writing in English. When you write a non-gender inclusive term, e.g. businessman or businesswoman, WProofreader marks it and gently suggests unbiased alternatives like business person.

Another case is swear words. Profanity is an absolute no-no-no (ok, sometimes there may be solid reasons). So, we’ve also created custom rules to detect profanity in writing.

Try Style guide now

Advanced user-level custom dictionaries

Dictionaries are like universes in alphabetical order.

For several months, we’ve worked hard on improving the user-level dictionary functionality, which is one of the most demanding features among our clientele. The newest updates include:

the complete refactoring of user custom dictionary functionality aimed at more consistent behavior and better security
unified functionality of user and company custom dictionaries
a new option to add/delete entire word lists (earlier, this was possible only on a word level)
a case sensitivity approach for better recognition of dictionary words

Discover more about custom spell-check dictionaries in our blog.

Enhanced spelling and grammar engine

Languages are ever-developing.

Although we’re working proactively in the direction of AI-based engine development, we do not abandon the maintenance of traditional rule-based engines. Another product update is our enhancements to the traditional spelling and grammar engines.

We’ve added and improved grammar and spelling rules for more than 10 languages, including Dutch, British English, German, French, Polish, Spanish, and Ukrainian. Plus, we’ve done some security enhancements by remediating vulnerabilities in the 3rd party components.

WebSpellChecker AI achievements

This spring was about big things in the WebSpellChecker AI realm:

a capability to recognize gibberish/obscure words in case the “enforceAI” option is enabled and one of the English languages is selected.
the “enforceAI” option for enhanced text checking for English dialects. Under the hood, it will replace the classic algorithmic engines with an AI-powered engine. You can test the difference on the demo page by enabling or disabling the Enforce AI-powered proofreading option.

Finally, we’re excited to announce that our NLP team has participated in the UNLP workshop (The Second Ukrainian Natural Language Processing Workshop), a part of the 17th EACL 2023, and introduced the custom RedPenNet approach for Grammatical Error Correction.

For Track 2 (GEC+Fluency), we got 0.6771 F0.5 score 💪 Congratulations and wishes are accepted!

Accessibility testing

The power of the web is in its universality.

Bearing this in mind, we stay dedicated to develop the solutions which comply with the global accessibility standards. Recently, we’ve accomplished the Voluntary Product Accessibility Template (VPAT) report, available on demand.It helped us to understand the level of accessibility provided by the WProofreader core and planned the updates and improvements of the product. The VPAT report is accessible per request.