Skip to content

AI Integration & MCP

Create Express Forge v4 is built to be "AI-Native". This means every project you scaffold comes with built-in infrastructure to help AI coding assistants (like Cursor, Claude, or Antigravity) understand and work with your codebase more effectively.

Model Context Protocol (MCP)

We provide a dedicated MCP server that allows AI tools to interface directly with your project's documentation and scaffolding logic.

Using the MCP Server

You can add our MCP server to your favorite AI IDE (like Cursor) or agent:

  • Server Name: Create Express Forge MCP
  • Command: npx -y @create-express-forge/mcp

Once connected, your AI assistant can:

  • Fetch the latest documentation for any feature.
  • Explain the project architecture.
  • Help you generate new modules following the project's specific patterns.

LLM Context Files

Every scaffolded project includes a set of .txt and .json files in the docs/public (or equivalent) directory that are specifically optimized for Large Language Models.

llms.txt

Located at the root of your documentation, this file provides a high-level summary of the project, its tech stack, and key patterns. AI assistants use this as a "map" to understand your code.

ai.json

A machine-readable manifest that describes the project's features, available commands, and documentation structure.

How it Helps

  1. Hallucination-Free Code: By providing the AI with direct access to our documentation via MCP, it is less likely to suggest deprecated APIs or incorrect patterns.
  2. Instant Onboarding: New developers using AI tools can understand the entire architecture just by asking the assistant to "summarize the project structure using the MCP server".
  3. Automated Documentation: Our build process automatically keeps these LLM context files in sync with your source code.

Configuration

You can customize the AI behavior by modifying the ai.json file in your project root or by updating the MCP server settings.

PRO TIP

If you are using Cursor, add https://create-express-forge.js.org/llms.txt to your @docs to give the AI agent full knowledge of the framework!

Released under the MIT License.