PostHog MCP Server

What the PostHog MCP server does

How models use it and what it is built for.

The server covers most of what a product team does inside PostHog. Analytics is the core: an agent can run HogQL — PostHog’s SQL dialect over the event schema — or ask for an insight to be generated from a natural-language description, then read the result. It can list and create dashboards and insights, so an exploration that proves useful can be saved rather than repeated. Feature flags are fully managed: list them, read their definitions and rollout conditions, create new ones and update existing ones, which is what makes this practical for shipping rather than only reporting. Error tracking exposes recent errors and their details, experiments and surveys are reachable, and a documentation search tool answers questions about PostHog itself. Authentication is OAuth — the first connection opens a browser and the credentials are reused afterwards — and PostHog runs the server in both US and EU regions, so EU Cloud data stays in the EU.

Tools the PostHog MCP server exposes

Typical tools an AI model can call. Exact names vary by version.

• query-run — execute HogQL against your event data
• get-sql-insight — turn a natural-language question into a query and run it
• insights and dashboards — list, read and create saved analytics
• feature flags — list, read, create and update flags and their rollout conditions
• error tracking — list recent errors and inspect the details of one
• experiments and surveys — read and manage running experiments and surveys
• docs-search — search PostHog documentation from inside the conversation

Example prompts to try

Copy any of these into MCP Agent Studio after connecting.

• What is the signup-to-activation conversion rate this month versus last?

• Which errors started appearing after the release on Tuesday?

• Create a feature flag for the new checkout, rolled out to 5% of users.

• Write a HogQL query for weekly retention by acquisition channel and run it.

Models on MCP Playground

This is not a single-model product: you get the same MCP connection with 60+ models (Claude, GPT, Gemini, DeepSeek, open-weight, and more), you can switch mid-conversation, and you can open Compare mode to run the same prompt against multiple models at once. The card above is a suggested starting point for this server — not the only choice.

Check an AI agent can actually use the PostHog MCP server

Listing tools proves the server is reachable, not that a model can work with it. Evals go further: they read every tool on the server, write a test suite from its real schemas, and run it — code decides pass/fail on the responses (schema conformance, error codes, pagination, result caps) while a scoring model grades plain-English tasks driven through the tools.

PostHog MCP server — FAQ

Common questions about connecting, scoping and using it safely.