Wordgard: In-browser rich-text editor from the creator of ProseMirror

Original Article ↗ Hacker News Discussion ↗

React and architectural considerations

  • Several comments note ProseMirror is awkward in React because React wants to “own” the DOM and use pure-state updates, while ProseMirror/Wordgard manage DOM more directly.
  • Wordgard’s transaction-focused update model (view receives transactions, not just state) is seen as a potential improvement for precise DOM updates and integration with UI frameworks, though docs could better address React interop.

Motivation, relation to ProseMirror, and switching cost

  • People ask why Wordgard exists instead of a ProseMirror v2 and whether it’s worth migrating existing setups (e.g., TipTap-based).
  • The answer from the author in-thread: ProseMirror will continue to be maintained; Wordgard embodies new design insights that couldn’t be cleanly added in a backward-compatible way.
  • There is no simple upgrade path; switching likely involves significant work and may not be worthwhile if ProseMirror already meets needs.

Technical design details

  • Wordgard’s edit model is compared to “retain/keep + action” sequences used in some collaborative editors, praised as easier to reason about than ProseMirror’s steps.
  • It uses a custom selection layer instead of browser selection, which some see as saner for complex editing behavior.
  • Zero external dependencies apart from small utility packages by the same author.
  • Mentions/“@”-style completions and other utilities are candidates for inclusion but not yet clearly defined.

Comparison with other editors and the web platform

  • Multiple comparisons with Lexical: some praise Lexical’s power and plugin system and active maintenance; others distrust Meta as steward or prefer the independent project.
  • There is broad frustration that, after decades, browsers still lack a robust standard rich-text element; contenteditable is viewed as buggy and inconsistent, making robust editors a huge effort.
  • Older contenteditable-based editors (TinyMCE/FCKEditor era) are contrasted with modern model-driven editors like ProseMirror/Wordgard/Lexical.

Mobile behavior and stability

  • Several users report serious issues on mobile (Android and iOS): undo not working, selection bugs, full-selection deletion breaking the editor, autocorrect/suggestion replacements deleting words, swipe-backspace not working.
  • Others confirm some bugs, and at least one specific issue was filed and fixed.
  • Consensus: architecture is promising, but mobile behavior is still unstable and needs real-world testing.

Ecosystem, hosting, and maintenance concerns

  • Some worry about ProseMirror’s future as many major products rely on it; the thread clarifies it will continue to be maintained and is MIT-licensed, so it can be forked.
  • Brief concern that the self-hosted code repo was down; others confirm it works and note motivations for moving off GitHub (uptime, sanctions risk, product direction).

UX, art, and general sentiment

  • Strong positive reactions to the site’s hand-drawn artwork and non-AI illustrations.
  • Mixed feelings on WYSIWYG in general; some prefer markdown and emphasize autosaving drafts (e.g., via localStorage) as crucial product behavior rather than a technical limitation.