PostHog MCP server

PostHog × Claude Code

Use the PostHog MCP server
in Claude Code

Query analytics, insights, and feature flags in PostHog. This guide connects PostHog to Claude Code (the terminal agent) through a governed gateway — one URL that also works in every other MCP client you use, with your rules and a full log attached.

Live-probed · Security-scanned · OAuth sign-in

Setup: PostHog in Claude Code

Claude Code adds MCP servers via the `claude mcp add` command. The gateway route takes about two minutes:

  1. Connect PostHog to your free gate gateway from the verified directory.
  2. In your terminal: claude mcp add gate https://gateturbo.com/api/mcp --transport http
  3. Approve the sign-in when Claude Code opens it in your browser.
  4. PostHog's tools now appear in every Claude Code session — with per-tool allow/ask/block rules if you want them.

What to ask once it’s connected

How did signups trend this week?
Which feature flags are currently active?
Show the funnel from landing page to signup.
Which pages convert best?

Why route it through a gateway

One URL, every client

The same PostHog connection works in Claude, ChatGPT, Cursor, VS Code too — no per-client re-wiring.

Rules before regret

Allow, ask, or block every tool PostHog exposes — reads can flow freely while writes require your approval.

Every call logged

A plain-language activity log of everything Claude Code did with PostHog — searchable, exportable, auditable.

Scanned and probed

PostHog passed a live handshake probe and a security scan before listing — and drift detection watches it after you connect.

Also available: PostHog in Claude · PostHog in ChatGPT · PostHog in Cursor · PostHog in VS Code