Meet ArchyPress MCP: AI-Driven Publishing From Your Code Editor

What if publishing a blog post was as easy as writing a commit message — and your AI agent handled the rest?

Content publishing has always lived in a separate world from where developers spend their time: the code editor. You write features in VS Code, then context-switch to a CMS dashboard, manually paste text, upload images, fiddle with formatting, and click Publish. It works. But it's slow, disconnected, and frankly beneath you.

Today we're launching ArchyPress MCP — a Model Context Protocol server that brings the entire publishing pipeline into your IDE. Draft blog posts. Generate AI images and video. Compose rich editorial layouts. Publish — all without leaving your code editor.

What Is MCP, and Why Should You Care?

The Model Context Protocol is an open standard that lets AI agents — like GitHub Copilot — interact with external tools through a structured, typed interface. Think of it as a USB-C port for AI: one standard connector, unlimited capabilities.

ArchyPress MCP exposes 40+ tools and 3 intelligent prompts that give any MCP-capable agent full read-write access to your publishing workflow. No API keys to juggle manually. No brittle shell scripts. Just natural language instructions, and your agent takes care of the rest.

How It Works: The Pipeline

Here's the end-to-end flow when you ask your AI agent to create and publish a blog post through ArchyPress MCP:

1. Your agent calls create_draft with a title, slug, categories, and SEO metadata

2. It generates AI images (or captures Playwright screenshots) and uploads them as assets

3. It composes the body using set_body_from_blocks — structured blocks like headings, quotes, card sets, images, CTAs

4. It attaches a feature image and optionally a feature video

5. One final call to publish, and the post is live

The entire flow happens over MCP tool calls. Your agent never touches raw HTML. The server validates every block against its schema before writing the TipTap document to the database.

The Full Toolbelt: 40 Tools, 3 Prompts

ArchyPress MCP ships with a comprehensive set of tools across two domains — blog posts and website pages. Here's a breakdown:

Rich Content, Not Just Paragraphs

One of the biggest limitations of traditional AI-generated blog content is that it reads (and looks) like a wall of text. ArchyPress MCP solves this fundamentally. The block composer supports 12 distinct block types:

• Headings (H1–H3) with inline formatting and links

• Paragraphs with rich inline marks — bold, italic, hyperlinks

• Pull quotes for editorial hooks and key insights

• Bullet and ordered lists

• Images with alt text — screenshots, AI-generated art, mermaid diagrams

• Videos — AI-generated clips or uploaded MP4s

• Embeds — YouTube, Vimeo, and more

• Image sliders — carousels for multi-screenshot walkthroughs

• Cards — standalone highlight cards with images and links

• Card sets — grids for feature comparisons, key points, attributes

• CTAs — call-to-action banners with primary and secondary buttons

• Mermaid diagrams — rendered as images via mermaid.ink

This post itself was written entirely through MCP tool calls — including the diagram above, the card grid, this very paragraph, and the CTA at the bottom. No CMS dashboard was opened. No manual copy-paste.

Website Pages: Not Just Blog Posts

MCP isn't limited to blog content. ArchyPress MCP includes a full page builder with 30+ section variants. Your agent can generate an entire landing page from a simple prompt:

From hero sections with parallax effects to pricing tables, FAQ accordions, testimonial walls, and contact forms — every section is a TiptapJSON node with typed attributes. The AI page outliner proposes the structure; the section generator fills in the content; your agent composes and publishes.

Getting Started in 3 Minutes

Setting up ArchyPress MCP is straightforward. You need an MCP-capable AI client (VS Code with Copilot, Claude Desktop, Cursor, etc.) and an ArchyPress API key.

1. Install the @archy-ai/blog-post-mcp package from npm

2. Add your ARCHY_BLOG_API_KEY and ARCHY_BLOG_BASE_URL to your MCP configuration

3. Point your MCP client at the server — that's it

Once connected, your agent can immediately start calling tools. Ask it to draft a post, generate illustrations, compose a rich editorial layout, and publish — all in one conversation.

The Three Built-In Prompts

Beyond raw tools, ArchyPress MCP ships with three intelligent prompts that orchestrate complex multi-step workflows:

Why This Matters

Publishing should be a natural extension of building — not a separate chore you dread after the code is done.

For developer-led teams that ship product blogs, changelogs, landing pages, and marketing content alongside their code, MCP-driven publishing eliminates the context switch entirely. Write code, then ask your agent to write the post about it. Review, tweak, publish. Ship.

The structured block system means your AI-generated content isn't just text — it's a rich, editorial layout with cards, pull quotes, diagrams, and CTAs that matches what a human designer would produce. And because everything flows through a typed MCP schema, the output is always valid, always structured, always ready to render.

What's Next

ArchyPress MCP v0.0.5 ships with 40 tools and 3 prompts covering the full blog and page lifecycle. We're actively building toward:

• Scheduled publishing — set a publish date and let MCP handle the timing

• Multi-workspace support — manage several sites from one MCP connection

• Template library — pre-built page and post templates for common patterns

• Analytics integration — pull in performance data to inform content strategy

• Collaborative review — submit drafts for team review before publishing