Claude Is Now Connected to My Portfolio via MCP

Something quietly remarkable happened today: Claude — Anthropic's AI assistant — is now directly connected to this portfolio through a custom Model Context Protocol (MCP) server hosted at txz.cool.

What Is MCP?

Model Context Protocol (MCP) is an open standard that allows AI models like Claude to connect to external tools and data sources in a structured, reliable way. Instead of copy-pasting content back and forth, Claude can read, create, update, and manage content directly — with full awareness of what's already there.

What the MCP Server Supports

The MCP integration exposes full CRUD control over every content type in this portfolio. Here's exactly what it can do:

Content Types

Each type has its own data shape:

Blog — title, body, cover image, tags
Project — title, body, summary, cover image, tags, and external links (with labels and URLs)
Experience — title, body, organization, start/end date (YYYY-MM), cover image, tags
Photography — title, body, cover image, tags, and a pinned flag to highlight featured shots

Available Operations

List contents — browse all items with filters by type (blog, project, experience, photography) and status (draft or published), search by title, and paginate results
Get content — fetch full details of a single item by ID
Create content — create a new item (always saved as draft first)
Update content — edit any field of an existing item without affecting others
Publish — make a draft item publicly visible
Unpublish — revert a published item back to draft
Delete — permanently remove a content item

Why This Is Interesting

Most people think of AI assistants as conversational tools — you ask a question, you get an answer. But MCP pushes that boundary further. Claude isn't just responding to prompts here; it's acting as an intelligent agent with real write access to a live system.

This opens up some genuinely useful workflows: asking Claude to summarize a project and publish it, keeping a photography log updated automatically, or even having Claude draft experience entries based on a resume.

A Small but Meaningful Step

This is a small integration, but it represents something I find exciting about where AI tooling is heading — toward systems that are genuinely connected to the things that matter to you, not just trained on generic knowledge.

If you're curious about building something similar, the MCP spec is open and surprisingly approachable. And if you'd like to see what Claude does next with this portfolio, well — stay tuned.