PostHog MCP ServerUnclaimed

The PostHog MCP server is a hosted endpoint that lets an AI agent work with PostHog using plain text requests. It works with any MCP-compatible client, including PostHog Desktop, Claude Code, Claude Desktop, Cursor, Codex, VS Code, Windsurf and Zed.

Once connected, an agent can read and write across PostHog's products. The page gives examples such as listing the top errors in a project for the week, pulling the stack trace for a recent crash and proposing a fix, marking an issue as resolved, and assigning future TypeErrors to a team. It also lists analytics queries, feature flag and experiment management, SQL, CDP destinations and support ticket triage.

Setup is done either with the PostHog Wizard, which runs as npx @posthog/wizard mcp add and installs the server into supported editors, or manually by pointing a client at the server URL https://mcp.posthog.com/mcp. The authentication server routes the account to the US or EU data region based on the login used.

The page says that connecting to the MCP server and calling its tools is free. Some tools use LLMs internally and may be billed as PostHog AI spend, and those tools only work when AI data processing is enabled in the organisation settings.

The server is aimed at developers and product teams who already use PostHog and want to query their data or manage flags and errors from a code editor or AI assistant rather than switching to the browser. The page also links to the server's source code on GitHub, a Use cases page with example prompts, and a reference of every tool the server exposes.

PostHog's wider pricing applies to the products the MCP server touches. The pricing page describes a free tier that resets every month across products, with pay-as-you-go billing above those limits and billing limits that can be set per product.