Connect Looker to AI: Let Claude and Cursor Read Dashboards, Explores, and Metrics

Why Connect Looker to AI?

Looker is where teams define metrics, build dashboards, and explore data. When you need a number mid-conversation, you open Looker, find the right dashboard or explore, filter it, read the result, switch back. For teams with dozens of dashboards and hundreds of explores, finding the right metric interrupts flow.

With an MCP server pointed at Looker, AI agents can query your BI layer directly. Check dashboard values, browse explores, pull metric definitions without leaving your terminal.

What AI Agents Can Do With Looker Access

Once connected, Claude, Cursor, or Codex can:

• Read dashboards — "What's the current MRR on the revenue dashboard?"
• Browse explores — "What dimensions are available in the orders explore?"
• Pull metrics — "What's the 7-day active user count?"
• Check definitions — "How is the churn rate metric defined in LookML?"
• Filter results — "What's revenue by region for Q1?"
• List content — "What dashboards exist in the Executive folder?"

How It Works

1. Open DataFaucet and enter your Looker instance URL
2. Browse Looker — click through dashboards, explores, looks
3. DataFaucet captures the read operations as MCP tools
4. Deploy — your AI agents now have structured access to query Looker

You're browsing the Looker web interface, not sharing API credentials. DataFaucet captures HTTP interactions from the web UI. Your Looker permissions stay in your browser session.

Example Queries

"What dashboards are in the Sales folder?"
"What's the current value on the daily active users tile?"
"What dimensions and measures are in the orders explore?"
"How is the net revenue retention metric defined?"
"What's the conversion rate from the marketing dashboard?"

When This Saves Time

• Standups — pull key metrics without opening Looker
• Code review — verify metric definitions match implementation
• Incident response — check dashboard values from terminal
• Planning — query business metrics during architecture discussions
• Reporting — grab numbers for Slack threads without context-switching

Security Model

DataFaucet connects through the Looker web interface. Access is scoped to whatever your Looker permissions allow. If you only have access to certain models and explores, the MCP server has the same constraints. No API keys or client secrets stored separately.

Getting Started

Open Looker in your browser, copy the instance URL. Paste it into DataFaucet, browse through dashboards and explores. Deploy. Your agents can now query BI metrics from your terminal.

Create your Looker MCP server →