Looker MCP is a Model Context Protocol server that gives AI agents access to your Looker instance. It provides functionality to run queries, build and render dashboards, organize content, schedule delivery, and set alerts through a standardized endpoint that works with MCP clients like Claude, Cursor, and Gumloop.

Is Looker MCP free to use?

Gumloop has a free plan that includes access to MCP integrations, including Looker. Paid pricing starts at $37/month with higher usage limits and additional features.

Do I need Looker credentials?

Yes. Looker uses API3 client-credential authentication, so you’ll create a service-account client ID and secret in Looker Admin > Platform > API and add them to Gumloop along with your API host URL. Gumloop stores them securely and mints a short-lived token for each call, so you don’t have to put credentials in environment variables, manage tokens, or write any code.

What AI clients work with the Gumloop Looker MCP server?

Any MCP client that supports the Model Context Protocol. That includes Claude Desktop, Cursor, Gemini-based tools, and Gumloop’s AI agents. You can also connect it through frameworks like LangChain if you’re building custom AI agents.

Can I combine Looker MCP with other integrations?

Yes. With AI agents in Gumloop, you can chain Looker with any other MCP integrations in a single agent. For example, an agent can run a Looker query, render the matching dashboard to PDF, save it to Google Drive, and post a summary to Slack.

Is my Looker data secure?

The Looker MCP server respects your existing Looker access controls and permissions, so an AI agent can only reach what your connected service account can. Your API host URL is validated before each call, and binary exports such as PDFs and images are saved to Gumloop storage and returned as a reference, scoped to your workspace when you set that option. For details on Gumloop’s security practices, see trust.gumloop.com.

Can the Looker MCP write data, or is it read-only?

It does both. The server can read your content and data (run queries, list and fetch dashboards, Looks, folders, boards, models, scheduled plans, and alerts) and it can write: creating and updating dashboards, tiles, filters, Looks, folders, boards, scheduled plans, and alerts, and moving or copying content between folders. All queries themselves are read-only against the warehouse, so the agent never inserts, updates, or deletes database rows.

What can’t the Looker MCP do?

It cannot edit LookML or touch Git and development mode, run raw SQL through SQL Runner, manage users, groups, roles, or permissions, manage database connections, generate embed or SSO tokens, or read System Activity and usage logs. It focuses on querying your data and managing the content built on top of it.

We raised a $50M Series B led by Benchmark

Looker MCP Server

Name: Looker MCP Server
Author: Gumloop

Connect to the Looker MCP server to run queries, build dashboards, render reports, schedule delivery, and set alerts using AI agents on Gumloop, Claude, or Cursor.

Talk to Sales

Try Looker now

Type what you want done. Sign in and run it live with an AI agent.

Looker

Installation

Set up the Looker MCP server in Gumloop

Do this once to provision your hosted server URL.

Create a Gumloop account

To use this MCP, you need a Gumloop account. If you don't have one yet, you can create one for free.

Add and authorize the Looker server

In Gumloop, open Connectors and add Looker. Paste your Looker API key when prompted. The credential is stored securely in Gumloop.

Then use it in your client

Use in Gumloop

Use Looker in an agent

Once Looker is set up, just open any Gumloop agent, add Looker as a connector, and start chatting with the agent.

Tools (67)

List Dashboards
Search dashboards in Looker.
Get Dashboard
Get a dashboard by ID.
Create Dashboard
Create a dashboard.
Update Dashboard
Update a dashboard.
Delete Dashboard
Delete a dashboard.
Move Or Copy Content
Move or copy a dashboard or Look to a folder.
List Dashboard Elements
Search dashboard tiles.
Create Dashboard Element
Create a dashboard tile.
Update Dashboard Element
Update a dashboard tile.
Delete Dashboard Element
Delete a dashboard tile.
Create Dashboard Filter
Create a dashboard filter.
Update Dashboard Filter
Update a dashboard filter.
Delete Dashboard Filter
Delete a dashboard filter.
Search Content
Search Looks, dashboards, and LookML dashboards.
List Looks
Search Looks.
Get Look
Get a Look by ID.
Create Look
Create a Look.
Update Look
Update a Look.
Delete Look
Delete a Look.
List Folders
Search folders.
Get Folder
Get a folder by ID.
Create Folder
Create a folder.
Update Folder
Update a folder.
Delete Folder
Delete a folder and its child content.
List Content Favorites
Search content favorites.
Create Content Favorite
Favorite a Looker content item.
Delete Content Favorite
Remove a content favorite.
Get Content Metadata
Get content metadata by ID.
Update Content Metadata
Update content metadata.
Validate Content
Run Looker content validation.
List Boards
List boards.
Get Board
Get a board by ID.
Create Board
Create a board.
Update Board
Update a board.
Delete Board
Delete a board.
Create Board Section
Create a board section.
Update Board Section
Update a board section.
Delete Board Section
Delete a board section.
Create Board Item
Create a board item.
Update Board Item
Update a board item.
Delete Board Item
Delete a board item.
Run Look
Run a saved Look and return results.
Run Query
Run a saved query and return results.
Run Inline Query
Run an inline Looker query and return results.
Create Query
Create a reusable Looker query.
Get Query
Get a query by ID.
Create Query Task
Create an async query task.
Get Query Task
Get async query task status.
Get Query Task Results
Get async query task results.
Create Dashboard Render Task
Create an async dashboard render task.
Create Look Render Task
Create an async Look render task.
Get Render Task
Get async render task status.
Get Render Task Results
Get async render task results (the rendered PDF/image).
List Lookml Models
List LookML models.
Get Lookml Model
Get a LookML model.
Get Lookml Model Explore
Get a LookML explore.
List Scheduled Plans
List scheduled delivery plans.
Get Scheduled Plan
Get a scheduled delivery plan.
Create Scheduled Plan
Create a scheduled delivery plan. Set exactly one target entity: dashboard_id, look_id, lookml_dashboard_id, or query_id.
Update Scheduled Plan
Update a scheduled delivery plan; include only the fields to change. The target entity (dashboard_id, look_id, lookml_dashboard_id, query_id) is mutually exclusive — set at most one.
Delete Scheduled Plan
Delete a scheduled delivery plan.
Run Scheduled Plan Once
Run a scheduled delivery plan once.
List Alerts
Search alerts.
Get Alert
Get an alert by ID.
Create Alert
Create an alert.
Update Alert
Update an alert.
Delete Alert
Delete an alert.

What is Looker MCP?

The Looker MCP server gives AI agents access to your Looker instance. That means agents can run ad hoc queries against any LookML model and explore, run saved Looks and queries, introspect the data model to learn what is queryable, build and edit dashboards with their tiles and filters, render dashboards and Looks as PDF or images, organize folders and boards, schedule automated deliveries, and set threshold alerts. It works across the full Looker domain, from dashboards and Looks to LookML models, scheduled plans, and alerts.

If your team spends time pulling the same numbers by hand, rebuilding report packs every Monday, or copying dashboard exports into emails and decks, an AI agent can take over a lot of that repetitive work. Describe what you need, and your AI agent will handle the querying, rendering, and delivery for you.

MCP stands for Model Context Protocol. It is an open standard that gives AI agents a way to connect to external tools and services. Instead of writing code against the Looker API3 endpoints, logging in to mint access tokens, and parsing the JSON yourself, you add your Looker API host URL and client credentials to Gumloop once. After that, you can query and report on your data just by chatting with your AI agent.

Related MCP servers

Tableau MCP Server

Interact with Tableau to access views, workbooks, datasources, and Pulse metrics

Google BigQuery MCP Server

Query and manage BigQuery datasets, tables, and execute SQL

Snowflake MCP Server

Interact with Snowflake data warehouse for executing SQL queries, managing tables/databases, and accessing data

Databricks MCP Server

Manage Databricks clusters, jobs, and job runs

ClickHouse MCP Server

Query ClickHouse Cloud services via SQL, manage services, backups, ClickPipes and costs, and manage ClickStack observability (dashboards, alerts, sources)

Sigma Computing MCP Server

Interact with Sigma Computing to manage workbooks, members, teams, workspaces, and data connections

What you can do with Looker MCP on Gumloop

Run ad hoc queries against your data model
Compose queries on the fly with run_inline_query, specifying the model, explore, fields, filters, sorts, and pivots without a pre-saved Look. This lets an agent answer data questions directly from the warehouse and feed the results into the rest of its work.
Run saved Looks and queries
Execute any saved Look or query and get the data back in the format you want: JSON, CSV, SQL, HTML, Markdown, XLSX, or rendered images. The agent picks the format that fits the next step, whether that is analysis or a report.
Introspect the LookML data model
Browse LookML models and explores with get_lookml_model_explore to discover available dimensions, measures, and drill fields. An agent can learn what is queryable and then auto-generate accurate inline queries against it.
Build and manage dashboards
Create, update, and delete dashboards, add and edit tiles backed by queries, and configure dashboard filters. The agent can also move or copy dashboards and Looks between folders to keep content organized.
Render dashboards and Looks as files
Kick off async render tasks to produce PDF, PNG, or JPG exports with control over paper size, orientation, and layout. Rendered files are saved to Gumloop storage and returned as a reference, and setting the workspace option scopes them to your Gumloop workspace for downstream use.
Organize and audit content
Manage folders and boards with their sections and items, search across Looks and dashboards, favorite content, and run content validation to catch broken references and modeling errors before they reach stakeholders.
Schedule automated delivery
Create and manage scheduled plans that send dashboards, Looks, or query results to email, webhook, S3, or SFTP on a cron or datagroup trigger. The agent can also run a plan once for an ad hoc send.
Set threshold alerts
Configure alerts on dashboard tiles with comparison logic, a cron schedule, and multiple notification destinations, so the team hears about a metric crossing a line without watching a dashboard.

How to connect the Gumloop Looker MCP Server

1
Create a free Gumloop account
Sign up at gumloop.com. No credit card required.
2
Add the Looker MCP server
Copy your MCP server URL from Gumloop and add it to your preferred client (Claude, Cursor, or Gumloop workflows). You'll authorize on first use.
3
Start using Looker in your AI workflows
That's it. Your AI agent can now run queries, build and render dashboards, schedule delivery, and set alerts across your Looker instance. Use it inside a Gumloop automation, in Claude Desktop, or in Cursor.

Looker MCP use cases

Self-serve metrics for business stakeholders

A RevOps lead asks the agent for last quarter’s pipeline by region. The agent introspects the LookML model, runs an inline query against the right explore, and returns the numbers in chat. It can post the same answer to Slack so the wider team sees it without opening Looker.

Automated executive report packs for BI teams

Every Monday an agent renders a set of dashboards to PDF, pulls the headline figures with saved queries, and emails the pack to leadership. Because rendering runs as an async task, the agent can build several exports and collect them once they are ready.

Dashboard build-out for analytics engineers

An analytics engineer describes a new dashboard in plain terms. The agent creates the dashboard, adds tiles backed by inline or saved queries, configures the filters, and files it in the right folder, turning a tedious clicking session into a single request.

Metric monitoring for operations teams

An ops team wants to know when refund rate crosses a threshold. The agent sets up an alert on the relevant dashboard tile with a cron schedule and notification destinations, then chains a Slack message so the on-call owner gets pinged the moment it trips.

Content governance for Looker admins

A Looker admin asks the agent to audit a cluttered instance. The agent searches across dashboards and Looks, runs content validation to surface broken references, and reorganizes folders and boards so stakeholders can find the reports that matter.

Why use Gumloop for Looker MCP

Skip the token plumbing
Add your Looker API host URL and API3 client ID and secret once, stored securely. No env vars, token minting, or coding against the Looker API necessary.
Works with multiple MCP clients
Claude Desktop, Cursor, or inside Gumloop. Same server URL, works with any MCP client.
Chain Looker with 100+ other integrations
An agent can run a Looker query, summarize the result, write it to Google Sheets, and post a recap to Slack, all in one run.
Enterprise-grade and scalable
Built for teams, with role-based access controls and dedicated support for Pro users. For details on Gumloop’s security practices, see trust.gumloop.com.
Pricing includes a free plan
Test on the free tier. Paid plans start at $37/month with higher usage limits and additional features.

Frequently asked questions

Ship Looker agents in minutes

Connect any AI agent to 100+ MCP servers, zero setup.

Talk to Sales

Looker MCP Server

Try Looker now

Installation

Set up the Looker MCP server in Gumloop

Create a Gumloop account

Add and authorize the Looker server

Then use it in your client

Use Looker in an agent

Tools (67)

List Dashboards

Get Dashboard

Create Dashboard

Update Dashboard

Delete Dashboard

Move Or Copy Content

List Dashboard Elements

Create Dashboard Element

Update Dashboard Element

Delete Dashboard Element

Create Dashboard Filter

Update Dashboard Filter

Delete Dashboard Filter

Search Content

List Looks

Get Look

Create Look

Update Look

Delete Look

List Folders

Get Folder

Create Folder

Update Folder

Delete Folder

List Content Favorites

Create Content Favorite

Delete Content Favorite

Get Content Metadata

Update Content Metadata

Validate Content

List Boards

Get Board

Create Board

Update Board

Delete Board

Create Board Section

Update Board Section

Delete Board Section

Create Board Item

Update Board Item

Delete Board Item

Run Look

Run Query

Run Inline Query

Create Query

Get Query

Create Query Task

Get Query Task

Get Query Task Results

Create Dashboard Render Task

Create Look Render Task

Get Render Task

Get Render Task Results

List Lookml Models

Get Lookml Model

Get Lookml Model Explore

List Scheduled Plans

Get Scheduled Plan

Create Scheduled Plan

Update Scheduled Plan

Delete Scheduled Plan

Run Scheduled Plan Once

List Alerts

Get Alert

Create Alert

Update Alert

Delete Alert

What is Looker MCP?

Related MCP servers

What you can do with Looker MCP on Gumloop

How to connect the Gumloop Looker MCP Server