Comparisons

Every editor on this page is a good piece of software. They are not interchangeable, because they are not solving the same problem: some are frameworks you build an editor with, and some are editors you install.

Blockleaf is the second kind, built on the first kind.

At a glanceCopy Anchor Link

TiptapCopy Anchor Link

Choose Tiptap when you want to own the decisions. It is the headless ProseMirror toolkit Blockleaf is built on, it is MIT-licensed, and it is maintained at a pace nothing else here matches.

The cost is that an editor is not what you get. You get extensions, and you assemble the schema, the menus, the shortcuts, the paste handling, the upload lifecycle and the styling yourself. That is the correct design for a toolkit. It is simply a lot of work to repeat, and repeating it is what Blockleaf exists to stop.

Collaboration, comments, AI, version history and document conversion are features of Tiptap's hosted platform rather than the MIT package, and start at $59 a month. If you need those, Tiptap's own stack is the shortest route to them and Blockleaf has no answer.

NovelCopy Anchor Link

Choose Novel if you like it as it stands today. It is an Apache-2.0 Notion-style editor on Tiptap, it got the feel right earlier than anyone, and the code is pleasant to read.

It has stopped moving. novel@1.0.2 was published in January 2025 and the repository has had no pushes since. For a dependency that owns your content surface, that matters: unmaintained is a decision you inherit, not one you make. Blockleaf covers much the same ground and intends to keep shipping.

BlockNoteCopy Anchor Link

Choose BlockNote if its block model fits your product. It is the closest thing on this list to a direct alternative — a Notion-style block editor, also on ProseMirror, under active development, with collaboration and an AI extension already in the box.

Two differences worth weighing. It is MPL-2.0 rather than MIT, which is permissive but file-level copyleft, so check it against your own licence policy. And it ships its own block document format and its own component styling; Blockleaf's bet is a smaller, explicitly versioned schema and a stylesheet of plain custom properties that imposes nothing on your build.

Editor.jsCopy Anchor Link

Choose Editor.js if you want clean block JSON without React. It is framework-agnostic, Apache-2.0, has a large plugin ecosystem, and its output is about as tidy as block JSON gets.

It is not a ProseMirror editor, and it shows in the things ProseMirror is good at: rich inline marks, selection across blocks, input rules, and collaborative editing primitives. If your content is mostly structured blocks with light inline formatting, that trade is fine. If people will be writing prose in it all day, it is less fine.

LexicalCopy Anchor Link

Choose Lexical if you are building an editor, not installing one. It is Meta's editor framework — MIT, fast, well-tested, framework-agnostic with React bindings.

Like Tiptap, it is a framework: the comparison is Lexical versus Tiptap, and then whatever you build on top versus Blockleaf. If you are comparing against Blockleaf directly, you are comparing a weekend of integration against several weeks of construction.

Where Blockleaf is weakerCopy Anchor Link

For symmetry, the honest list:

• No collaboration. No CRDT, no presence, no sync. BlockNote and Tiptap's platform both have it.
• Young. 0.0.5, with the schema still pre-0.1.0. The versioning promise is real, but the track record behind it is short.
• React only. The core is deliberately framework-free so that other bindings stay possible, but today React is the only one shipped.
• Smaller plugin ecosystem. It is Tiptap's ecosystem underneath, which helps a great deal — but there is no catalogue of Blockleaf-specific blocks to shop from.

Where it is strongerCopy Anchor Link

• The stored shape is a contract, versioned and migrated rather than whatever the editor happened to emit. See the document.
• Reading is a separate package. @blockleaf/renderer renders stored JSON on a server with no editor and no ProseMirror — the read path of a content site does not ship an editor at all. See server rendering.
• No CSS framework, ever. One stylesheet, --bl-* custom properties, no prose class and no build step. See the style guide.
• The expensive parts are opt-in. Syntax highlighting sits behind its own entry point. See code blocks.
• Bundle size is a budget, not a hope. The published bundle is measured in CI against a hard limit and the build fails when it grows past it.

We have not measured anyone else's bundle for this page, and you should not trust a number we did not produce. Measure the editor you are considering with the extension set you will actually ship.