Get Started
Connect Your AI Environment
Set up the Webiny MCP server so your AI coding agent can assist with Webiny development.
- what the Webiny MCP server provides to your AI coding agent
- how to connect it using the built-in setup command
- how to configure it manually for any supported agent
Overview
Webiny ships with a local MCP (Model Context Protocol)
server that gives your AI coding agent structured, up-to-date knowledge about Webiny APIs, extension patterns, and project conventions. Instead of relying on general training data, the MCP server provides detailed skills — focused guides on specific topics like content models, lifecycle events, dependency injection, and infrastructure — directly to the agent when it needs them.
The server exposes two tools your agent calls automatically:
list_webiny_skills— returns a catalog of all available skills with names and descriptionsget_webiny_skill— loads the full documentation for a specific skill by name
Connecting the MCP server early means your agent can assist with all the customization work that follows.
Quick Setup
The fastest way to connect is to run the built-in setup command from your Webiny project root:
Replace claude with your agent of choice:
claude— Claude Codecursor— Cursorwindsurf— Windsurfcopilot— GitHub Copilot / VS Codecline— Cline
This command creates two files:
- MCP config file — tells your agent how to start the Webiny MCP server
- Instruction file — tells your agent to use the Webiny tools when working on Webiny-related tasks
Once done, restart your agent or reload the editor window. The Webiny MCP server becomes available immediately.
Manual Setup
If your agent is not covered by the built-in adapters, or you prefer to configure things yourself, follow the two steps below.
Register the MCP Server
Add the following entry to your agent’s MCP configuration file:
The location of this file depends on your agent:
| Agent | Config file |
|---|---|
| Claude Code | .mcp.json (project root) |
| Cursor | .cursor/mcp.json (project root) |
| Windsurf | .windsurf/mcp.json (project root) |
| GitHub Copilot / VS Code | .vscode/mcp.json (project root) |
| Cline | .vscode/cline_mcp_settings.json |
| Zed | ~/.config/zed/settings.json (under context_servers) |
| Claude Desktop | ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) |
GitHub Copilot expects servers instead of mcpServers in its config file. The rest of the structure is the same.
Add Agent Instructions
Add the following to your agent’s instruction or rules file so it knows to use the Webiny tools:
The location of the instruction file depends on your agent:
| Agent | Instruction file |
|---|---|
| Claude Code | CLAUDE.md (project root) |
| Cursor | .cursor/rules/*.mdc |
| Windsurf | .windsurf/rules/*.md |
| GitHub Copilot / VS Code | .github/copilot-instructions.md |
| Cline | System prompt in settings |
| Zed | System prompt in settings |
Verify the Connection
The simplest way to verify is to ask your agent a question like:
“List all available Webiny skills.”
If the MCP server is connected, the agent will call list_webiny_skills and return a catalog of topics like content-models, lifecycle-events, dependency-injection, and others.
You can also test the server independently using the MCP Inspector:
Connect, click List Tools, and confirm you see list_webiny_skills and get_webiny_skill.
A separate remote MCP server is also available at https://mcp.docs.webiny.com/mcp for tools that are not project-based IDEs — such as Claude Desktop, ChatGPT, or Claude.ai. This server provides access to the full Webiny documentation rather than the project-level skills described on this page.
What's Next
With your AI environment connected, you’re ready to make your first customization with AI assistance.
Continue to First Customization.