# Looker MCP > The Looker MCP gives your AI agent full control over enterprise business intelligence and data analytics. Instead of clicking through dashboards or writing complex SQL, you talk to your data. This connector lets you list dashboards, execute queries against specific models, audit saved reports, and search content metadata—all via natural conversation. ## Overview - **Category:** industry-titans - **Price:** Free - **Tags:** bi-platform, data-visualization, dashboard-orchestration, data-modeling, sql-queries, enterprise-analytics ## Description Getting deep insights from a large BI platform usually means getting stuck in the UI, running into permission walls, or writing complex SQL that takes forever to debug. Not anymore. This MCP connects your AI agent directly to your Looker instance, letting you manage and query data through plain English chat. You can ask your agent to list all managed dashboards and get their detailed configuration metrics instantly. Need a specific number? You run an inline query against any model or view just by asking for the dimensions and measures. Furthermore, if you need historical context, the agent accesses saved 'Looks,' pulling out model mappings and filters so everyone sees consistent data reports. It's like having a specialized data analyst sitting next to you, ready to answer complex questions about your organization’s metrics. Accessing this power through Vinkius makes it easy for any MCP-compatible client to connect, giving you immediate control over your entire BI environment. ## Tools ### list_dashboards Lists all available dashboards in your Looker instance. ### get_dashboard Retrieves the complete details and query mapping for a specific dashboard ID. ### list_looks Lists all saved data mappings (Looks) that track specific datasets. ### get_look Retrieves the full mapped details for a specified target Look object. ### list_folders Lists root folders to analyze explicit environment structures within your tenant. ### search_content Searches content metadata across the entire instance for specific mapping targets. ### run_inline_query Executes queries by building models to fetch literal dimensions and measures dynamically. ## Prompt Examples **Prompt:** ``` List the last 5 dashboards created in my Looker instance ``` **Response:** ``` I've retrieved the 5 most recent dashboards. Highlights include 'Q1 Sales Overview' (ID: 123), 'User Engagement Deep-Dive' (ID: 456), and 'Inventory Health' (ID: 789). Would you like to see the filters for the sales overview? ``` **Prompt:** ``` Run a query using model 'sales' and view 'orders' for fields 'orders.created_date' and 'orders.total_amount' ``` **Response:** ``` Executing inline query… I've retrieved the data from the 'sales' model. For the last 10 orders, the created dates range from yesterday to today, with a total amount of $12,450. Would you like the full 100-row result set? ``` **Prompt:** ``` Find all dashboards related to 'Marketing ROI' ``` **Response:** ``` I've searched your content metadata and identified 2 dashboards: 'Global Marketing ROI' (ID: 987) and 'Channel Specific ROI - Beta' (ID: 654). I can fetch the UUIDs and folder paths for either of these. ``` ## Capabilities ### Audit and Inventory Data Assets List all available dashboards and search through content metadata to locate specific datasets or analytical assets. ### Run Real-Time Data Queries Execute custom queries against defined models and views, fetching literal dimensions and measures on demand. ### Inspect Saved Reports (Looks) Retrieve full details about saved reports to understand the exact model mappings and filters used for consistent reporting. ### Manage Folder Structure Enumerate root folders and top-level models, allowing you to audit permissions and organizational structure across your entire Looker tenant. ## Use Cases ### Finding the Right Report After a Merger A business user needs a report on 'Global Marketing ROI' but doesn't know which folder it lives in. They ask their agent, and it uses `search_content` to locate all relevant dashboards and provides the UUIDs and path for them. ### Auditing Dashboard Dependencies A platform engineer suspects a dashboard is pulling stale data. Instead of manually checking every tab, they ask their agent to use `get_dashboard` on the ID, instantly retrieving the detailed query structure for review. ### Quick Ad-Hoc Number Check A data analyst needs to know the total revenue from 'orders' last quarter but doesn't want to write a full SQL script. They prompt their agent to `run_inline_query` on the sales model and get the summary immediately. ### Reviewing Historical Metrics A manager needs proof of Q1 performance metrics. Instead of asking a colleague who built it, they ask the agent to use `get_look` for the saved 'Q1 Sales Overview,' confirming exactly which filters and models were applied. ## Benefits - Get instant visibility into your entire BI environment. Instead of clicking through multiple menus to find out what dashboards exist, you simply ask the agent to list all available dashboards using `list_dashboards`. - Run complex data checks without writing SQL. You can execute queries against specific models and views just by asking for the dimensions and measures you need, leveraging `run_inline_query` for real-time results. - Maintain reporting consistency. If you need to know how a number was calculated months ago, you don't have to hunt through old files. The agent accesses saved 'Looks' using `list_looks` and provides the full mapping details. - Understand your data architecture quickly. Need to audit permissions or see what content exists? Use `list_folders` to enumerate root folders and understand the organizational structure across all environments. - Pinpoint exactly where information lives. When you can’t find a report, use `search_content`. This tool searches metadata across the instance so your agent knows where to look for that key dataset. ## How It Works The bottom line is you talk to your BI platform instead of navigating its interface. 1. Subscribe to this MCP and enter your specific Looker Base URL, Client ID, and Client Secret. 2. Connect your preferred AI client (Claude, Cursor, etc.) to the Vinkius catalog. 3. Ask your agent a question in natural language; it uses the tools to fetch data or metadata from Looker. ## Frequently Asked Questions **How do I find out what dashboards exist in my Looker instance using the Looker MCP?** You use the `list_dashboards` tool. This instantly gives your agent a comprehensive list of every dashboard available, so you don't waste time guessing which ones are active. **Can I query data without knowing the specific SQL? Looker MCP?** Yes. You use `run_inline_query`. Instead of writing code, you just tell your agent which model and which dimensions or measures you want, and it builds the query for you. **What is 'Looks' in Looker and how does the MCP help me with them?** A 'Look' is a saved report detailing specific data mappings. The MCP lets you use `list_looks` and `get_look` to retrieve these details, ensuring consistent reporting across your organization. **If I can’t find a dashboard, what tool should I use with the Looker MCP?** Use `search_content`. This searches the metadata across your entire instance. It's much better than just listing folders because it actively hunts for content based on keywords. **Does this MCP handle complex folder structures in Looker?** Yes, you can use `list_folders` to enumerate the root folders and top-level models. This helps audit permissions and understand the full organizational structure of your tenant.