Lexical libraries comparison

How @rozie-ui/lexical compares to the existing per-framework Lexical wrappers. Lexical's editor core is framework-agnostic and mounts anywhere; every wrapper exists only to glue reactive state, register plugins, and bridge decorator nodes. The result is the thinnest binding ecosystem of any major editor: a first-party React package, a stale Vue and Solid story, a single-maintainer Svelte package, and nothing for Angular or Lit. Rozie delivers the same idiomatic editor component on all six targets, including the two the ecosystem leaves entirely unserved, Angular and Lit, as pre-compiled per-framework packages.

Research snapshot: 2026-08-10. Versions and download counts move; treat them as of that date.

The wrappers at a glance

Wrapper	Package	Maintainer	Status
React (official)	@lexical/react	Meta (first-party)	Healthy — the reference binding
Vue (community)	lexical-vue	wobsoriano	Stale — pinned many minors behind core
Solid (community)	lexical-solid	mosheduminer	Stale — tracks core slowly
Svelte (community)	svelte-lexical	umaranis	Capable but bus-factor-1; actively published yet pins core months behind (0.40 vs 0.49)
Angular	—	—	No maintained wrapper exists
Lit	—	—	No wrapper exists
Rozie	@rozie-ui/lexical-*	One Learning Community	Same API on all 6 targets (incl. Angular + Lit)

The case for Rozie is strongest for Angular and Lit (no wrapper at all) and Vue / Solid (stale, pinned behind core); even the actively-published Svelte package pins its @lexical/* dependencies several minors behind the current core. The recurring failure mode across the ecosystem, a thin React transliteration pinned N versions behind, is exactly what Rozie counters: every target ships from the same pinned engine, so no framework's package lags the others.

Feature matrix

Capability	@lexical/react	lexical-vue	lexical-solid	svelte-lexical	Angular (none)	Lit (none)	@rozie-ui/lexical
Mount editor	✅	✅	✅	✅	hand-roll	hand-roll	✅
Shared editor context for children	✅	✅	✅	✅	hand-roll	hand-roll	✅ $inject('rozie-lexical-editor')
RichText baseline	✅	✅	✅	✅	hand-roll	hand-roll	✅ (shell + RichTextPlugin)
History (undo/redo)	✅	✅	✅	✅	hand-roll	hand-roll	✅ HistoryPlugin
List plugin	✅	✅	✅	✅	hand-roll	hand-roll	✅ ListPlugin
Link plugin	✅	✅	✅	✅	hand-roll	hand-roll	✅ LinkPlugin
Selection-reading toolbar	build it yourself	build it yourself	build it yourself	✅ (rich sample)	hand-roll	hand-roll	✅ bidirectional Toolbar
Decorator-node component renderer	✅	✅	✅	✅	hand-roll	hand-roll	✅ neutral-descriptor bridge (6 targets)
Tracks the latest Lexical core	✅	⚠️ stale	⚠️ stale	⚠️ pins behind	—	—	✅ engine pinned in lockstep
Same API on all 6 targets	❌	❌	❌	❌	❌	❌	✅

Where Rozie wins today

• First-class packages everywhere, including the two the ecosystem leaves entirely unserved: Angular (no wrapper) and Lit (no wrapper). An Angular or Lit developer today hand-rolls the whole editor integration; Rozie gives them a standalone component / web component.
• A shared-editor context ($inject('rozie-lexical-editor')) with the same shape on every target: the extension seam custom plugins and the toolbar ride on, versus each framework's bespoke context/provider idiom.
• A bidirectional selection-reading toolbar out of the box: buttons that both dispatch commands and reflect the caret's current formatting, where React/Vue/Solid consumers build it themselves.
• A neutral-descriptor decorator bridge proven end-to-end across six targets by the reference @mention node (including Lit through an open shadow root), plus an authoring recipe for custom nodes.
• Lockstep-pinned engine. lexical and every @lexical/* subpackage are pinned to the same core version in every package, so the family never drifts behind core the way the community Vue/Solid/Svelte wrappers have.

Staging — what ships today vs v1.1

Every target (including Lit) ships today; the staging table — what ships now versus what remains deferred to v1.1 (Markdown-shortcuts + Tables) — lives on the showcase page's roadmap.

Scope limits

• All six targets ship, but Lit carries a browser floor. The Lit build requires Chrome 137+ / FF 142+ / Safari 17+ (getComposedRanges for cross-shadow selection); the other five have no such floor. The comparison rows for Angular/Lit describe the upstream ecosystem (no wrapper); Rozie ships both.
• @rozie-ui/lexical is pre-1.0. The surface (the shell + four plugins + toolbar + one decorator node) is stable and gate-verified, but the package is younger than the multi-year @lexical/react.
• Deliberately unopinionated. The toolbar and chips ship minimally styled by design (theme + your CSS is the styling path). This is the family ethos, not an omission; see the scope & posture notes.
• No collaboration / no SSR hydration. @lexical/yjs collaboration and server-rendered-document hydration are out of scope (the editor is import-safe under SSR, but there is no hydration path).

Cross-references

• Lexical — showcase & API — the full @rozie-ui/lexical surface, composition, and recipes.
• LexicalEditor.rozie source
• Decorator node authoring recipe — the neutral-descriptor + per-target bridge pattern.