looker-mcp

# Looker MCP Server A Model Context Protocol (MCP) server for Looker. This server dynamically provides **all 483 Looker API methods** as tools for your favorite MCP client! ## Quick Start (Automatic Setup) If you are using an MCP client (like Claude Desktop, Cursor, or Cline), you can run this server directly from GitHub without any manual setup! The client will automatically fetch and install everything. Add the following to your client's `mcp_servers.json` (or equivalent configuration): ```json { "mcpServers": { "looker": { "command": "uvx", "args": [ "--from", "git+https://github.com/yuvaraj-97/looker-mcp.git", "looker-mcp" ], "env": { "LOOKER_BASE_URL": "https://mycompany.looker.com:19999", "LOOKER_CLIENT_ID": "your_client_id", "LOOKER_CLIENT_SECRET": "your_client_secret" } } } } ``` *Note: This utilizes [`uv`](https://docs.astral.sh/uv/) for zero-install execution. If you do not have `uv` installed, you have a few alternatives:* **Alternative 1: Using pipx** Change the command to `pipx` and the args to: `["run", "--spec", "git+https://github.com/yuvaraj-97/looker-mcp.git", "looker-mcp"]` **Alternative 2: Using pip (Standard Python)** 1. Install the package globally by running: `pip install git+https://github.com/yuvaraj-97/looker-mcp.git` 2. In your JSON config, set the `command` to `looker-mcp` and leave `args` empty (`[]`). ## Example Usage Scenario Once the Looker MCP server is connected to your AI assistant, you can ask the AI to perform complex data workflows naturally. Here are some examples of what you can say: **Prompt 1: "Find the Q3 Sales dashboard and give me a summary of the data."** 1. The AI uses the `search_dashboards(title="Q3 Sales")` tool to get the dashboard ID. 2. The AI uses the `dashboard(dashboard_id="...")` tool to retrieve the dashboard details and underlying query IDs. 3. The AI runs `run_query(query_id="...", result_format="json")` t…