Run your email workflow
from Claude.
The Pagenflow MCP server connects Claude, Claude Code, Cursor, or any MCP-compatible agent to your account. Create projects, build and validate email templates, and send test emails in plain language, with no local install.
claude mcp add --transport http pagenflow https://mcp.pagenflow.workers.devClaude Code one-liner. Other clients are in the setup steps below.
- template_duplicateBlack Friday → Black Friday (Cyber Monday)
- template_field_opsHeadline, CTA label and hero image updated
- template_validate0 errors · 1 warning (missing alt text)
- template_test_sendSent to you@brand.com
Codex
KimiWhat you can do with the Pagenflow MCP server
The Model Context Protocol (MCP) is the open standard that lets an AI assistant call real apps. Pagenflow exposes your account as tools: say it in plain language and it happens, with exactly the permissions you already have.
Templates
Build, validate, publish and export (HTML, React, Shopify Liquid), send test emails, and generate preview screenshots.
Projects
Create, star, archive, duplicate and publish projects.
Teams & members
Invite teammates, change roles, switch workspace.
Design systems & fonts
Brand colors, typography tokens and logos, kept in sync.
Reusable components
Browse and manage the blocks your team shares.
Billing & credits
Check your plan and credit balance before you kick off a batch of exports or test sends.
Integrations & recipients
Review connected integrations and manage your test-send recipient pool.
Not covered yet: the drag-and-drop canvas itself, connecting a brand-new third-party integration, and completing a billing checkout. Those still need a browser session.
How to connect Pagenflow to your AI assistant
Three steps to a hosted MCP server: nothing to install or keep running.
- 1
Generate a personal access token
Go to Account Settings → API Tokens, name it (e.g. “Claude Desktop”) and copy it. It's shown only once, and grants the same access as signing in, so keep it only in your AI client's config.
- 2
Add Pagenflow as a remote MCP server
Pick your client and paste the same URL and token everywhere. It speaks standard Streamable HTTP.
Settings → Developer → Edit Config
JSON{ "mcpServers": { "pagenflow": { "url": "https://mcp.pagenflow.workers.dev", "headers": { "Authorization": "Bearer pfmcp_live_your_token_here" } } } } - 3
Ask your assistant to get to work
Try “list my Pagenflow teams” or “create a project called Onboarding Emails”. Your assistant discovers the tools on its own and only sees the ones your account may use.
Well-organized tools
Every tool is named and scoped by domain, so your assistant picks the right one every time.
Scoped to your account
A token can only do what its owner can do in the app. Revoke it any time from account settings.
Zero install
A hosted Streamable HTTP server. Point your client at a URL, with no local process to keep running.
Pagenflow MCP server FAQ
Everything you need to know about using Pagenflow through MCP.
What is the Model Context Protocol (MCP)?
MCP is an open standard that lets AI assistants like Claude connect securely to external tools and data sources. Instead of copy-pasting between an app and a chat window, an MCP-compatible assistant can call a defined set of "tools" directly — in Pagenflow's case, that means creating projects, managing templates, inviting teammates, and dozens of other actions.
Is my personal access token safe? Can I revoke it?
A personal access token grants the same access as signing in — treat it like a password. You generate and revoke tokens yourself from Account Settings → API Tokens, and revoking one takes effect immediately. We recommend giving each MCP client (e.g. "Claude Desktop", "Claude Code") its own token so you can revoke one without affecting the others.
Can an AI assistant edit the visual design of my templates?
Not yet. The MCP server covers everything around a template — creating, organizing, publishing, exporting, testing, and screenshotting it — but the drag-and-drop design itself still happens in Pagenflow's editor. Support for AI-driven visual editing is on our roadmap.
Which AI tools work with Pagenflow's MCP server?
Any MCP-compatible client: Claude Desktop, Claude Code, Cursor, Windsurf, Cline, Codex, Kimi, and other agent frameworks that support the Streamable HTTP MCP transport. You connect by adding our server's URL and your personal access token to the client's configuration — no local install required.
Do actions taken through MCP use my plan's credits?
Yes — an MCP tool call that would cost credits in the app (like exporting a template or sending a test email) costs the same via MCP. Read-only actions, like listing projects or checking your credit balance, are always free.
Give Claude the keys to your inbox work.
Generate a token, paste one command, and start managing your email templates in plain language. Revoke access any time.
claude mcp add --transport http pagenflow https://mcp.pagenflow.workers.devNo local install · Works with Claude, Cursor, Windsurf, Cline, Codex & Kimi