home / mcp / yamu mcp server

Yamu MCP Server

Configures a per-project MCP server entry to run the Yamu MCP server alongside your AI agents.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "keijiro-yamu": {
      "command": "node",
      "args": [
        "Library/PackageCache/jp.keijiro.yamu@(HASH)/Node/mcp-server.js"
      ]
    }
  }
}

You run a small, per-project MCP server that lets AI agents access local commands and data sources. This server runs alongside your project, so agents can execute tasks and retrieve results in the context of your codebase, making workflows faster and more automated.

How to use

Connect your MCP client to the per-project MCP server you configure for your project. With the server running, you can have your agent request the local command execution capability and query available data sources that you expose through the MCP surface. Use the client to start tasks, feed results back to the agent, and iterate on prompts and workflows. Make sure your agent has access to the per-project settings, and that the MCP server is started whenever you work in the project to ensure responses come from the correct environment.

How to install

Prerequisites you need before installing and running the MCP server:

- Node.js installed on your system. Check with node -v.

Step-by-step commands you can follow to prepare and configure the per-project MCP server in your project:

# 1) Verify Node.js is installed
node -v

# 2) Open the per-project agent settings file for your AI agent
#    Example paths depending on the agent:
#    Claude Code -> .mcp.json at project root
#    Gemini CLI -> .gemini/settings.json

# 3) Add or update the MCP server entry for the yamu server in the settings file.
#    You will replace (HASH) with the actual version hash found in your environment.
```"}]} ,{