PostHog
Open-source product analytics suite with Max AI and built-in LLM observability.
About
PostHog is an open-source, self-hostable product analytics platform that bundles product analytics, session replay, feature flags, A/B experiments, surveys, a data warehouse, error tracking, and LLM observability into one suite. Its Max AI assistant answers natural-language questions over analytics data, and its AI Observability product captures LLM generations with cost, latency, and quality breakdowns plus prompt management and evaluations.
Founded in 2020 in Y Combinator's W20 batch by James Hawkins and Tim Glaser, PostHog raised a $70 million Series D led by Stripe in June 2025 at a $920 million valuation, with $119 million total raised and 190,000+ customers. The core is MIT-licensed with roughly 40,000 GitHub stars as of October 2026; the cloud free tier covers 1 million events per month, and the company says 98% of customers pay nothing. Caveat: several products are cloud-only, and self-hosting at scale carries real infrastructure cost.
What it can do
- Product analytics with funnels, retention, and SQL insights
- Session replay, feature flags, and A/B experiments
- Max AI natural-language analytics assistant
- LLM observability with prompt management and evals
Setup
Sign up for PostHog Cloud or self-host via Docker; install the SDK snippet in your app.
Prerequisites
An app or site to instrument; a PostHog account for cloud use.
Use cases
Works with
Trust score — 9/10
Scored 9/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 60; Trending score 86/100 from researcher assessment.
★ 40,000 GitHub stars
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.