JVLink

The server queries a local SQLite database populated from JRA-VAN DataLab. You need an active JRA-VAN DataLab subscription and must run the companion jrvltsql tool to download race data and create the keiba.db file before the MCP server can do anything useful.

The project ships an interactive installer that clones the repo, resolves dependencies, searches for an existing keiba.db, and writes your MCP client config automatically. Run the command below in your terminal and follow the prompts.

curl -fsSL https://raw.githubusercontent.com/miyamamoto/jvlink-mcp-server/master/install.sh | bash

If you prefer a manual setup or are on Windows, clone the repo and install Python dependencies via uv. The server auto-installs remaining packages on first launch (takes 30–60 seconds).

git clone https://github.com/miyamamoto/jvlink-mcp-server.git
cd jvlink-mcp-server
pip install uv
uv sync

Add the server block to your Claude Desktop configuration file. Set DB_PATH to the absolute path of your keiba.db file. On macOS the config lives at ~/Library/Application Support/Claude/claude_desktop_config.json; on Windows at %APPDATA%\Claude\claude_desktop_config.json.

{
  "mcpServers": {
    "jvlink": {
      "command": "uv",
      "args": ["run", "--directory", "/path/to/jvlink-mcp-server", "python", "-m", "jvlink_mcp_server.server"],
      "env": {
        "DB_TYPE": "sqlite",
        "DB_PATH": "/path/to/keiba.db"
      }
    }
  }
}

If you use Claude Code in the terminal, register the server with a single command instead of editing JSON manually.

claude mcp add jvlink \
  -e DB_TYPE=sqlite \
  -e DB_PATH=/path/to/keiba.db \
  -- uv run --directory /path/to/jvlink-mcp-server python -m jvlink_mcp_server.server

Quit and relaunch Claude Desktop. On first start the server installs remaining Python packages. Once the MCP tools icon appears in the chat interface, the server is ready. Ask Claude a racing question to confirm.