I cannot output JSON-LD in Etch. The security layer strips it. JSON-LD is Schema.org structured data — by definition, a web standard. Etch’s own documentation promises “Web Standards” and “Full Empowerment — complete control over your code without limitations.” Except, apparently, that one. Meanwhile, AI that can copy a stranger’s layout for $0.07 just became the top priority on the roadmap. It wasn’t on the list at all in January.
When I bought into Etch early, the pitch included a phrase I’ve heard a lot since: era 4. A new era of web building. I’ve asked what that means specifically. I got an explanation. I still don’t know what it means. What I do know is that “a new era” is the most recycled phrase in software marketing, right after “game changer” and “we’re reimagining X.” It signals existence, not direction.
But the documentation is more specific than the marketing phrase. So let’s use that.
What Etch promises
Etch describes itself as the first Unified Visual Development Environment for WordPress. The key principles listed in their own documentation include:
Full Empowerment — Complete control over your code without limitations.
Web Standards — Work with Semantic HTML, CSS, PHP, and JavaScript as they were designed to work.
Crystal Clean Output — Professional-grade code that you’d be proud to write by hand.
These are not vague aspirations. They’re specific claims. I bought in based on claims like these, and the core experience — real CSS applied directly at the element level — actually delivers on them. It’s the best builder experience I’ve used for that specific thing.
But claims are worth checking against reality.
The reality
I cannot output JSON-LD in Etch. The security layer strips script-adjacent elements. JSON-LD is Schema.org structured data — it’s literally how the web is designed to communicate meaning to search engines. It is, by definition, a web standard.
“Web Standards — Work with Semantic HTML, CSS, PHP, and JavaScript as they were designed to work.”
Except that one.
“Full Empowerment — Complete control over your code without limitations.”
Except that limitation.
I’ve written about the workaround. There shouldn’t be a workaround. And when I raised the need for data attributes on JS script elements — a specific, non-exotic, legitimate use case — nobody in the community understood why that would be useful. I was an edge case.
The serious use cases are always edge cases.
The roadmap tells the real story
On January 1st, 2026, the Etch priority list looked like this: stability, pattern library, recipes, PHP authoring, proper undo/redo, facets, external API loops.
AI was not on the list. Not immediate priority, not high priority — not mentioned at all.
Etch has shipped consistently. Weekly updates, real progress. That’s not nothing — it’s actually one of the reasons I stayed invested.
But the Components API was promised before V1. Developers were told it would come after the first stable release. It’s still not there. I still can’t build native components programmatically. That’s been on the list longer than anything currently marked “immediate priority.”
And yet: Native AI just jumped from not on the list at all in January to the single top priority today. Three months. No gradual climb up the backlog — straight to the front.
What changed between January and now wasn’t the product requirements. What changed was the community got excited about a demo.
That’s not a roadmap decision. That’s a community excitement detector attached to a backlog.
On the AI celebration
The community this week is excited because the AI panel rebuilt a complex hero section from a screenshot for $0.07. Someone else’s design, replicated quickly, cheaply. The enthusiasm is genuine. The milestone is questionable.
I tested the same AI panel differently. I asked it to help me add JSON-LD structured data — Schema.org markup, the invisible layer that tells search engines what a page actually means.
It doesn’t know how to do that.
So the AI can copy a stranger’s layout. It cannot build semantic infrastructure for your own site. For $0.07 you get someone else’s visual design. The work that actually matters is still fully manual — and currently blocked by a security layer that contradicts the documentation’s own promises.
The excitement reminds me of the first weeks after ChatGPT launched. Everyone was amazed it produced text. Not that it produced good text, not correct text — just that something came out. Same energy. Same pattern.
The pattern I’ve seen before
Elementor was going to change everything. Then DIVI. Then Oxygen. Then Bricks. Each started with a sharp idea. Each ended up trying to replace everything adjacent to it — and became exactly what it replaced: bloated, proprietary, feature-heavy in the wrong places.
Etch now wants to replace SEO plugins. Replace ACF. Replace Metabox. Add AI. The original pitch — clean builder, real code, no bloat, full control — is quietly becoming one feature among many.
“It’s all the benefits of a traditional page builder with none of the downsides.”
The downside of traditional page builders was always the same thing: they started focused and ended up trying to be everything. I’m not sure Etch has avoided that. I’m not sure it’s still trying to.
What I actually want
Not a replacement for every plugin in the WordPress ecosystem. Not AI that copies designs.
Let me output JSON-LD. Let me add data attributes where I need them. Deliver the Components API that was promised. Build the CSS-at-element experience deeper, not wider. Keep the output clean. Keep the code mine.
Era 4 could mean something real. Right now it’s a heading in search of a document.
Update: I asked the team directly in the Developers Lounge. On the Components API sequencing, fair enough, groups, repeaters, and logic were prerequisites. That makes sense.
On why AI jumped to the top of the roadmap from nowhere: “Changes in the market and how people are wanting to work.”
Which is exactly the point.
