Blocks as the Missing Layer of the Web

The web has always been built for human-readable documents first. HTML gives you a paragraph here, an emphasized word there, and CSS decorates it all. But that structure is shallow. If you publish a book citation — title bolded, author named, publisher credited — a computer reading the page has no way to know a book is even being discussed.

This isn’t a new problem. Tim Berners-Lee described the Semantic Web in 1999 as a vision where machines can analyze content, links, and transactions across the web, enabling intelligent agents to handle day-to-day tasks. The idea was that publishers would mark up data with standards like RDF or JSON-LD, guided by vocabularies from schema.org.

It hasn’t happened at scale. The barrier is practical: adding semantic markup to a published page is extra work with no immediate payoff. Bloggers finish a post, hit publish, and don’t return to embed machine-readable metadata. So schema.org markup remains rare in the wild.

The project at blockprotocol.org wants to change that with a simple premise: semantic markup will only take hold if adding it is easier than omitting it. The cost has to be zero or negative.

Making Structure Effortless

Imagine writing a blog post and wanting to share a book recommendation. You type /book. A search box appears. You type the title, pick from an autocomplete list, and a formatted block drops into your post — complete with structured data attached. The same flow works for any data type:

  • Type /address, autocomplete a location, and insert a block with full address semantics behind the scenes.
  • The visual display can be anything — simple text, a map, even a localized map — but the underlying semantics stay consistent.
  • Browsers or other tools can then recognize the content as an address and offer relevant actions, like navigation.

This isn’t about two data types. You can imagine thousands: events, people, research papers, ingredients. But right now, nothing like this works across the web. Every major editing environment — WordPress, Notion, Trello, Mailchimp — has its own concept of blocks. None are interoperable. A block built for one platform cannot be used in another, and building a new block type means waiting on that vendor to ship it.

An Open Protocol for Interchangeable Blocks

The proposal is an open standard. Developers build blocks that conform to this specification. Editors and platforms also conform, so any block works anywhere. A writer or vendor who creates a great block for, say, book citations makes it available to everyone — not just users of one product.

The Block Protocol is positioned as fully free, open, and public. Public or open-source blocks are welcome; so are proprietary or commercial ones. There’s no restriction on who builds or distributes them.

Where the Project Stands

After roughly a year of design work, the protocol’s specification has been refined through practical testing. To make the idea concrete, the team has built a WordPress plugin that embeds Block Protocol blocks into posts just like native WordPress blocks.

This matters because WordPress powers roughly 43% of the web. A block developer targeting the protocol reaches that entire audience immediately, without building anything WordPress-specific — no PHP required. The plugin also serves as an easier starting point than writing a custom WordPress plugin for developers who just want one new block type.

The plugin will be free and widely available in February, alongside version 0.3 of the Block Protocol specification. Early access is available now via the project site.

Those interested can also ask questions on the project’s Discord server at blockprotocol.org/discord. Project updates and discussion also happen on Mastodon, where Joel Spolsky posts as @[email protected] — described as a human-scale alternative to algorithm-driven social feeds.