PostHog review: features, pricing, and AI capabilities

Platform overview

PostHog AI is an in-product analyst. It answers plain-English questions against the project's own events, persons and warehouse schema, writes the SQL and builds insights. It also acts: creating dashboards, surveys, alerts and feature flags, editing filters in the UI, and keeping project memory across conversations. It's available in the web app, inline throughout the UI, and in Slack in beta, and it can hand code changes to a connected repository and return a pull request.

Self-driving PostHog goes further. Background agents cluster signals from errors, session replays, support tickets and external tools into reports, then open PRs in a sandbox. A human still merges.

Agent access is first-class. A free hosted MCP endpoint at https://mcp.posthog.com/mcp exposes function-calling tools across every PostHog product, including full experiment lifecycle control and governed semantic-layer metrics. Some MCP tools call LLMs internally and are only available when AI data processing is enabled.