Get the plugin
WP MCP/Integrations/Page builders

Page builders integration

Thrive Architect AI agent tool reference

Thrive Architect stores a layout's HTML in post meta, with the part before a Read More split, the CSS it generates per element, and a plain-text copy in post_content; landing pages store the same rows under their template's suffix. WP MCP reads the layout as a tree with a path per element and writes one element operation or a whole layout, keeping the Read More part in step. Untouched elements keep their data-css ids so their styles still apply. Thrive's optimized per-page assets were computed for the old layout, so they are switched off and the page loads Thrive's full styles and scripts until it is optimized again, and the plain-text copy is rebuilt through Thrive when it is loaded. Rollback restores all of it together.

detect-builderget-builder-contentupdate-builder-content

In practice

How Thrive Architect work goes with an agent.

Thrive Architect pages are HTML, not a block or node tree: the editor saves markup into tve_updated_post, and a landing page stores it under a key with its template's suffix. WP MCP addresses elements by a dotted path into that markup and pins every write to a SHA-256 hash of the stored layout, so an agent that read a page before someone else edited it is refused instead of overwriting. Beyond one page, it can read Thrive's site-wide design data, user templates and element catalog, which lets an agent reuse existing styles instead of inventing new ones.

What global colours and button styles does Thrive have on this site?

Runs get-builder-content with scope design_system. It reads the global colours and gradients, the favourites the picker keeps and the global styles per element kind. This scope is read only.

Change the headline on the webinar landing page.

Runs get-builder-content for the tree and hash, then update-builder-content with operation update, the element's path and expected_hash. The post is snapshotted, so layout, CSS, Read More part and optimized-asset flags roll back together.

Save the pricing section as a Thrive template and reuse it on /offer.

Runs update-builder-content with save_template and the element's path, which creates a user template through Thrive and records it for undo. Then an add with element template on /offer, which copies the template's CSS rules in under fresh tve-u ids.

Publish the draft landing page now that I approved it.

Runs update-builder-content with operation publish. Only draft, pending, private or future pages can be published this way, an already published page is refused, and the status change is one snapshotted write.

What your AI can do

Real Thrive Architect operations, every write reversible.

  • Detect Thrive Architect pages and landing pages
  • Read a page's element tree with a path per element
  • Update, add, remove or move one element, or replace the whole layout
  • Keep the Read More excerpt in step
  • Keep edited elements styled by falling back to Thrive's full assets
  • Roll back layout, CSS, excerpt and optimized assets together

Example prompts

Example Thrive Architect prompts.

Which landing pages are built with Thrive Architect?

Change the headline and button text on the webinar landing page.

Remove the countdown element from /offer now that the sale ended.

Under the hood

The Thrive Architect calls, under the hood.

wpmcp, thrive-architect

get-builder-content({ post_id: 88 }), element tree with paths

update-builder-content({ post_id: 88, builder: "thrive", operation: "remove", path: "0.3" }), snapshotted, restorable

Safety and access

What can be undone on Thrive Architect, and what costs what.

Recoverable on Thrive Architect

  • Every page write is one post snapshot, and rollback-operation restores layout, CSS, Read More part, optimized assets and plain-text copy together.
  • restore puts back a Thrive revision of the page, chosen by the revision id the read returns, and is itself a snapshotted write.
  • save_template records the template it creates for undo.

Thrive Architect limits and opt-ins

  • Writes work only on a page already detected as Thrive. A first layout is made in the Thrive editor, which sets up the page's other rows.
  • A page with a layout needs the expected_hash from the read. A missing or stale hash writes nothing.
  • An edit that crosses the page's Read More split is refused, so make it on one side of the split at a time.
  • The design_system and catalog scopes are read only. WP MCP does not change global colours or styles.

Free or Pro: Thrive Architect support is Pro. detect-builder, get-builder-content and update-builder-content register only with a Pro licence, and they cover the other supported builders too. Free keeps the content tools for the post as a whole, such as the post fields and the rendered page HTML reader, but cannot read or change Thrive's layout rows. Page edits need an expected_hash from the read.

Part of WP MCP Pro

The Thrive Architect integration is part of WP MCP Pro.

One Pro licence unlocks the Thrive Architect tools and every other Pro integration on this site: Elementor deep editing, the other page builders, the forms adapter pack and more, 133 Pro tools in all. The free plugin still covers the rest of WordPress and the full snapshot and undo engine, so you can try the safety model before you buy.

Frequently asked

Thrive Architect MCP: questions, answered.

Why does an edited page load more CSS afterwards?

Thrive's asset optimization trims each page's stylesheet and scripts for its current layout. After an edit those trimmed assets are out of date, so WP MCP switches them off and Thrive serves its full assets until the page is optimized again. That keeps edited elements from rendering unstyled.

With Thrive Architect, can the agent create a new Thrive page from scratch?

No. Writes work only on a page already detected as Thrive, because a new layout is made in the Thrive editor, which also sets up the page's other rows. The agent can then edit the layout, add library templates and publish the page.

What happens if I edit the page in Thrive while the agent works?

The agent's next write carries the hash it read earlier, and the stored layout no longer matches it. WP MCP answers stale_expected_hash and writes nothing. The agent reads the page again, sees your change, and retries from the current layout.

Are landing pages handled differently from normal pages?

They store the same rows under a key with the template's suffix, and WP MCP picks the right key from the page's landing template automatically. The read reports which landing template the page uses, so the agent knows which kind of page it is editing.

With Thrive Architect, can the agent change Thrive's global colours or styles?

No. The design_system scope reads the global colours, gradients and styles, and the catalog scope reads element defaults, but both are read only. Changing a site-wide style is done in Thrive itself. The agent can still match existing styles when it edits a page.

Point your AI at Thrive Architect.