Custom Umbraco MCP for Multi-Site Content Operations

Custom Umbraco MCP to Run Our Marketing Sites Without Backend Logins

April 23, 2026

Eveliko built a custom Umbraco MCP that handles content authoring, preview, SEO and JSON-LD, navigation, content pipelines, and full-site onboarding from Claude, Cursor, and other MCP clients, with both OAuth and API Key authentication supported. More than 100 content pieces shipped across the Eveliko-operated portfolio in ten weeks, and long-tail search coverage is becoming a byproduct of the workflow rather than a standalone project.

Umbraco CMS MCP (Model Context Protocol) Claude Cursor GitHub Copilot AI Integrations

When Every Small Content Task Still Means a Backend Login

Backend logins are a tax on small, frequent content work.

Most marketing work is small and frequent, publishing a blog post, updating a service page, fixing metadata on one page, and the overhead of logging into the CMS, finding the right node, editing, previewing, and publishing each time adds up to hours a week that produce nothing a reader ever sees.

Different tools require different MCP schemes.

Claude connects to MCP servers over OAuth while Cursor and developer automations expect API keys, so a serious Umbraco MCP has to support both schemes as first-class options or one audience ends up forced through the wrong mechanism and builds workarounds.

Campaigns go wrong in the same places every time.

A new case study is not one page but a blog post, listing entry, navigation update, JSON-LD, and preview link all at once, and done by hand across a content tree something always gets missed, usually the navigation or the schema, and it is only noticed after it has been live for a day.

Onboarding a new site should not take an afternoon.

Every new client or vertical brand needs the same Umbraco scaffolding of root, templates, navigation, and correctly scoped content, which is mechanical work that should be a single tool call rather than an afternoon.

A Custom Umbraco MCP That Runs the Full Content Lifecycle

An Umbraco MCP built around how content actually gets made.

Eveliko built a custom Umbraco MCP that exposes the full content lifecycle as MCP tools, pages, blog posts, case studies, navigation, preview links, all invoked through natural language from Claude, Cursor, or any compliant client, without the editor opening the Umbraco backend for the majority of tasks.

Both major authentication schemes supported: OAuth and API Key.

Claude connects over OAuth and Cursor over API keys, both shipped as Umbraco plugins, so content operators and developer tooling can live on the same MCP without either audience being forced through the wrong mechanism.

Content pipelines turn campaigns into a single reviewable unit.

A case study, blog post, listing entry, navigation update, and SEO change can be stacked into one pipeline, previewed as a whole, and published atomically, so nothing goes live half-finished and navigation is never forgotten.

JSON-LD and SEO are first-class tools, not afterthoughts.

Organization, FAQPage, CollectionPage, BlogPosting, and Service schemas are updated alongside the content they describe, which is how eveliko.com's GEO score moved from the mid-40s into the mid-60s without any manual schema editing.

Site onboarding is one tool call.

Launching a new client, vertical, or brand site is a single duplicate-and-scope operation instead of an afternoon of manual cloning and navigation rewiring, producing consistent results every time.

Proof 1: volume. More than 100 content pieces across the portfolio in ten weeks.

Between February and April 2026, more than 100 content pieces were published through this MCP across eveliko.com and several other Eveliko-operated sites, more than Eveliko's previous marketing presence accumulated in over a decade, and backend logins became the exception rather than the routine.

Proof 2: visibility. Long-tail search coverage is a byproduct now, not a project.

With structured schema plus volume combined, eveliko.com now has pages and JSON-LD covering long-tail queries it never produced pages for before, SharePoint-to-Sitefinity migration partner searches, custom MCP server for Umbraco, and similar structured-intent queries, and early long-tail rankings are beginning to appear for ground the site did not cover at all before the MCP made those pages cheap to ship.