Skip to content
Documentation

Your EMCP guide

Answers from our docs

YOUR CHAT, YOUR CHOICE

Before we chat

MSR Builds processes your question and up to four recent messages on our server to answer from the EMCP docs. AI is optional and needs your separate permission.

The chat is kept in this page's memory, not saved as a conversation in our database or browser storage. Our hosting and security services process connection information, such as your IP address, to prevent abuse.

Cloudflare Turnstile checks your browser and connection to prevent automated abuse when you send a question. No login is required.

Please leave out personal, sensitive or confidential information, passwords and API keys.

You can withdraw consent and clear this page's chat at any time. This cannot undo a request already sent or erase provider/security records.

Chat privacy, retention & your rights ↗ · Privacy contact

Starting chat consents to this processing. You can browse our docs without using the chat.

Page Snapshot

New in v3.3.0. get-page-snapshot returns one normalized digest of a page so an agent can reason about it from a single call instead of chaining get-page-structure + get-global-settings + list-global-classes and reassembling the pieces itself. It’s the read foundation the other v3.3.0 features build on.

Read-only, edit_posts, always-on, enabled by default.

emcp-tools/get-page-snapshot

Section titled “emcp-tools/get-page-snapshot”

Input:

  • post_id (required)
  • include (optional), heavy opt-in audit sections: ["performance", "a11y", "seo"]. Default [].
  • sections (optional): subset the core sections to control token cost. Default: all.
  • fresh (optional): bypass the heavy-section cache.

Core sections (free, computed in-process):

SectionWhat it contains
postid, title, slug, status, type, url, builder (elementor/gutenberg/classic), elementor_version
structurenormalized tree (id, kind, widget_type, label, depth, children) + counts (by_widget_type, max_depth, total_elements)
tokensglobal colors / typography / g- classes actually in use, each with a usage count, plus raw fonts_in_use / colors_in_use
responsivewhich elements carry per-device (tablet/mobile) overrides + rollup counts
contentheading outline, word count, image count, images-missing-alt, link/button counts
seo_liteh1 count, meta title/description, canonical, og:image (Yoast / Rank Math / core)
warningsstructural smells (empty containers, excessive nesting, missing H1)

Opt-in heavy sections (only when named in include):

  • performance: page-level score/grade + top findings from the Performance Analyzer. Free (needs manage_options).
  • a11y: accessibility audit summary. Pro (returns pro_gated: true on free builds).
  • seo: deep SEO audit summary. Pro (returns pro_gated: true on free builds).

Heavy sections are transient-cached (~15 min); pass fresh: true to recompute.

Why it matters

Section titled “Why it matters”

The snapshot replaces the parse-and-reassemble dance an agent otherwise does across four+ read tools. It’s cheaper (one call, light by default), and it’s the substrate the AI reasons over before editing, feeding cleanly into content search (reuse) and change history (undo from EMCP Tools → Safety → History).

1. get-page-snapshot          → understand the whole page at once
2. search-content "hero"      → find something similar to reuse
3. update-widget / build-page → make the change
4. list-changes / History     → review & undo if needed