MCP Support

The Model Context Protocol (MCP) allows you to extend Kodezi CLI with custom tools, external utilities, internal systems, or workflow-specific logic, without modifying the CLI itself. Once configured, MCP servers provide new tools that the AI can automatically detect and use.

Features

Here’s what MCP brings to Kodezi CLI:

• Dynamic Tool Loading: Tools from MCP servers become instantly available as soon as the server connects.
• Multiple Server Support: You can connect several MCP servers at the same time, and all their tools are merged into the CLI toolset.
• Stdio Based Communication: MCP uses simple standard input/output (stdio) for communication, making integration lightweight, portable, and platform-agnostic.
• Auto Discovery: Kodezi automatically detects and registers tools exposed by an MCP server—no manual configuration once the server is running.

MCP Configuration

Add your MCP servers in the kodezi.json configuration file.

{
  "mcp": {
    "database": {
      "type": "stdio",
      "command": "npx",
      "args": ["@kodezi/mcp-database"]
    },
    "filesystem": {
      "type": "stdio",
      "command": "node",
      "args": ["/path/to/mcp-server.js"]
    }
  }
}

What each field means

MCP servers automatically register the tools they provide once they are running.

Benefits of MCP Support

MCP brings powerful advantages to your workflow:

• Extend Functionality Without Changing the CLI: You can add new tools or behaviors simply by running an MCP server.
• Create Your Own Tools: Build capabilities tailored to your project, company, or automation pipeline.
• Team Wide Shareability Teams can share MCP servers to enforce standards or provide internal utilities.
• Standardized Integration MCP uses a structured protocol designed for interoperability and long term compatibility.

Learn more about MCP setup →