Payload MCP

What is Payload MCP?

Payload MCP is a Model Context Protocol (MCP) server for Payload CMS that enables AI assistants to help you develop, manage, and extend your Payload CMS projects. It provides tools for generating code, validating components, scaffolding projects, and performing specialized queries—all accessible to AI agents through the standardized Model Context Protocol.

How to use Payload MCP?

Install the package globally (npm install -g @ngyngcphu/payload-mcp), then configure your MCP client (e.g., Cursor) by adding the server entry to your mcp.json with npx @ngyngcphu/payload-mcp as the command. The server starts automatically when needed; optionally start it manually with npx @ngyngcphu/payload-mcp.

Key features of Payload MCP

• Generates boilerplate code for collections, fields, hooks, and more
• Validates Payload CMS components against best practices
• Scaffolds new Payload CMS projects with recommended structure
• Provides specialized queries for codebase insights
• Works with any AI assistant that supports MCP
• Runs entirely locally for privacy and security

Use cases of Payload MCP

• Rapidly scaffold a new Payload CMS blog project with authentication, collections, and admin customization
• Ask an AI assistant to generate a complete collection configuration with fields, hooks, and access control
• Validate existing Payload CMS components to catch issues early in development
• Add features (like authentication or deployment setup) to an existing Payload project via AI guidance
• Query and analyze a Payload codebase to get recommendations and formatted responses

FAQ from Payload MCP

What runtime environments does Payload MCP require?

Node.js v16 or higher and npm (Bun recommended). The server runs locally and communicates over HTTP.

How do I connect my AI assistant to Payload MCP?

Configure your MCP client (e.g., Cursor) by adding the server entry to its mcp.json with the command npx @ngyngcphu/payload-mcp. The server listens on port 3001 by default.

Can I use Payload MCP without an AI assistant?

Yes—you can start the server manually with npx @ngyngcphu/payload-mcp and interact with tools programmatically, but the primary design is to provide tools to AI agents via the Model Context Protocol.

What types of code can Payload MCP generate?

It generates boilerplate for collections, fields, config files, access control rules, hooks, custom endpoints, plugins, blocks, migrations, and React admin components.

Is Payload MCP ready for production use?

No—it is marked as a work in progress and not yet feature-complete. APIs, configurations, and CLI commands may change without notice. Use at your own risk.