looker-mcp-gemini-enterprise

Configure Looker MCP server integration with Gemini Enterprise via OAuth and PKCE.

Updated Jul 7, 2026

Search Tags:#integration #oauth #mcp #pkce #looker #lookml #gemini enterprise

Ready to run

npx skills add https://github.com/ricardolui/gcp-custom-agent-skills --skill looker-mcp-gemini-enterprise-ricardolui

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: looker-mcp-gemini-enterprise
Source: https://github.com/ricardolui/gcp-custom-agent-skills/tree/main/looker-mcp-gemini-enterprise
Command: npx skills add https://github.com/ricardolui/gcp-custom-agent-skills --skill looker-mcp-gemini-enterprise-ricardolui

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Connecting Looker's semantic layer to Gemini Enterprise is error-prone: OAuth registration, PKCE configuration, role permissions, and the failing Conversational Analytics (A2A) approach cause confusion and infinite processing loops. This Skill provides a validated, step-by-step path to a working MCP connector. ## Core Features & Use Cases - Looker-side setup: BigQuery connection, LookML model publishing, Roles/Model Sets/Permission Sets, OAuth client registration, and MCP tool activation. - Gemini Enterprise configuration: Exact MCP Data Store parameters including Authorization URL, Token URL, scopes, and mandatory PKCE verification. - Troubleshooting guidance: Documents why the Conversational Analytics A2A approach loops infinitely and recommends the MCP connector instead. - Use Case: An analytics engineer enables natural-language queries over LookML Explores in Gemini Enterprise chat and builds no-code agents grounded in governed business metrics. ## Quick Start Guide me through connecting my Looker instance to Gemini Enterprise using the MCP connector with OAuth and PKCE.

Frequently Asked Questions about looker-mcp-gemini-enterprise

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I connect Looker to Gemini Enterprise?▼

Register an OAuth client in Looker's API Explorer with the Gemini redirect URI, enable MCP tools under Admin > MCP, then create an MCP Data Store in Gemini Enterprise using the Looker /mcp, /auth, and /api/token endpoints with PKCE enabled.

Looker MCP vs Conversational Analytics API for Gemini Enterprise?▼

The MCP connector works successfully, while the Conversational Analytics (A2A) approach authenticates but enters an infinite processing loop without returning answers. The documentation recommends discontinuing A2A and using the standard Looker MCP connector.

Does Looker MCP require a client secret for OAuth?▼

No client secret is required. Looker's public OAuth flow uses PKCE, so the Client Secret field can be left empty or set to none, but PKCE verification must be enabled and the code_challenge_method=S256 parameter included.

Why are Looker fields missing from Gemini Enterprise responses?▼

Missing fields usually mean the LookML model was not published to production. Changes made only in Development Mode are invisible to the MCP server, so commit and deploy the model before querying.

What Looker roles does a user need for MCP access?▼

The user needs a Role combining a Model Set scoped to the target model and a Permission Set with explore permissions. Recommended assigned roles include Conversational Analytics User, Gemini, User, and Viewer.