🚀 Getting Started

This guide will walk you through installing, configuring, and running the Mcp-server.

Installation

The Mcp-server is a package within the swiftcode monorepo. When you install the dependencies for the entire project, it will be available.

# From the root of the swiftcode project
pnpm install

Client Configuration

To use the server, an MCP client (like an IDE extension) needs to know how to start it. This is typically done in the client's configuration file (e.g., settings.json).

Using NPX (Recommended)

This method allows the client to run the server using npx without needing a local installation.

{
  "mcpServers": {
    "swiftcode-npm": {
      "command": "npx",
      "args": ["-y", "@swiftcode/mcp"]
    }
  }
}

For Local Development

If you are developing swiftcode locally, you can configure the client to run the server directly from your build output.

{
  "mcpServers": {
    "swiftcode-local": {
      "command": "node",
      "args": ["/path/to/swiftcode/packages/Mcp-server/dist/index.js"]
    }
  }
}

WARNING

Remember to replace /path/to/swiftcode with the actual absolute path to your local project directory.

Using Docker

You can also build and run the server within a Docker container.

# 1. Build the Docker image
docker build -t swiftcode-mcp .

# 2. Run the container
docker run -it swiftcode-mcp

Starting the Server Manually

While typically started by a client, you can run the server manually for testing.

# The binary is named based on its package.json `bin` entry
swiftcode-mcp-server

The server will now wait for a client to connect over stdio. To stop it, press Ctrl+C.