PostHog

PostHog began as a self-hostable Mixpanel alternative and has since grown into a unified product intelligence platform covering event analytics, session recording, heatmaps, feature flags, and multi-variant A/B testing — all sharing the same underlying event model and user identity graph. The cloud version grants one million events per month free before usage-based charges begin, and the self-hosted path via Docker Compose or Kubernetes carries no event limits or feature paywalls. SDKs exist for web, iOS, Android, React Native, and a range of server-side languages.

For product and engineering teams, the main argument for PostHog is consolidation. A single deployment can replace the separate subscriptions you might otherwise hold with Google Analytics, Hotjar, LaunchDarkly, and portions of Segment. HogQL, PostHog's SQL interface, lets analysts query raw event tables directly without exporting to a data warehouse first. Native data warehouse connectors to BigQuery and Snowflake cover teams that want to join PostHog events against external business data.

The flip side is operational weight. Running PostHog at scale requires ClickHouse, Kafka, and Redis alongside the application containers — a materially heavier stack than Plausible or Umami. Cloud pricing also climbs faster than simpler tools for high-volume products once the free tier is exhausted. Teams that only need traffic counts and referrer data will find PostHog overbuilt; teams instrumenting a real product with feature experiments and session-level debugging will find it one of the most capable open-source options available.