MCP: Ask your AI about your users

The Model Context Protocol (MCP) is an open standard from Anthropic that lets AI clients connect to external data sources and tools. An analytics MCP server is a bridge: it translates your AI's requests into analytics queries and returns structured data the AI can reason over.

Without it, getting an answer from your analytics looks like this: open the dashboard, navigate to the right view, build or find the right chart, read the number, close the tab, context-switch back to what you were doing. If the answer leads to another question, repeat.

With OpenPanel's MCP server, you ask and get an answer—inside the tool you're already in.

"What's the 7-day retention for users who signed up last month?"
"Which funnel step has the biggest drop-off this week?"
"Show me the last 5 sessions for user@example.com"

OpenPanel hosts the server at api.openpanel.dev/mcp. There's nothing to install or run locally. You create an API client in your dashboard, generate a token, and add the URL to your AI client's config. That's it.

The server exposes 38 tools covering every layer of your product data:

• Events and sessions — query raw events, filter by property, inspect individual sessions
• User profiles — look up users, their event history, and activity timeline
• Funnels — run funnel queries for any defined event sequence
• Retention — get cohort retention by day, week, or month
• Page analytics — top pages, entry/exit pages, performance metrics
• Reports and dashboards — fetch saved report data by name
• Google Search Console — clicks, impressions, top queries, SEO opportunities (requires GSC connected)

How this compares to the Google Analytics MCP server:

Google Analytics MCP connects to GA4 data—pageviews, traffic sources, sessions in GA's data model. OpenPanel's MCP connects to your product data: custom events, user profiles, funnels you've defined, retention cohorts. If you're tracking what users do inside your product, OpenPanel's MCP answers those questions. GA4's doesn't.

Authentication uses a base64-encoded token built from your client ID and secret. The server is read-only—your AI can fetch and query data but cannot write events, modify users, or change any settings.