PostHog Review

Overview

PostHog is an open-source product analytics platform with 39,933 GitHub stars. One install covers event analytics, session replay, feature flags, A/B testing, surveys, error tracking, a data warehouse, data pipelines, and logs. The core code is MIT licensed, while the ee/ directory carries a separate enterprise license, so the openness is real but partial. The same product runs as PostHog Cloud in US and EU regions or self-hosted. Pricing is usage-based credits on top of a free tier that renews every month for every product: 1M events, 5K session recordings, 1M feature flag requests, 100K exceptions, 1,500 survey responses, 10GB of logs, and more, with no card required and usage stopping at the limits. Pay-as-you-go adds six projects, seven-year data retention, and unlimited usage. Credit add-ons were shown as USD 5, 25, and 20 packs in September 2026. The free tier is not a trial; the vendor says 97 percent of companies never pay. The AI features are part of the product rather than a bolt-on. PostHog AI answers questions about your data in plain language across web, Slack, desktop, and MCP, with 500 free credits each month, while AI Observability and Evaluations track LLM usage and Replay Vision assists session review. Integrations cover Slack, GitHub, Zapier, Segment, Sentry, and HubSpot. For marketing and product teams that want funnels, retention, and experiments in one place, PostHog replaces a stack of separate tools and keeps the free tier large enough for real work.

AI Capabilities

• PostHog AI assistant for natural language queries (500 free credits per month)
• AI Observability and LLM Evaluations
• Replay Vision for AI-assisted session replay review

Key Integrations

Best for

Product and growth teams that want analytics, session replay, feature flags, experiments, and surveys in one platform, starting free and scaling on usage.

Not for

Organizations that need a fully permissive open-source stack in every component, or teams with strict procurement rules that rule out a partially proprietary codebase and cloud-first roadmap.

Review notes

Assessed from posthog.com/pricing, the documentation, and the GitHub repository rather than a production deployment. The scope is the surprise: event analytics, session replay, feature flags, experiments, surveys, error tracking, a data warehouse, data pipelines, and logs all live in one product with one signup. Teams that would otherwise buy an analytics tool, a replay tool, a flag service, and a survey tool separately get one bill and one tag.

The pricing page is unusually blunt. Every product gets a free allowance that resets monthly with no card on file: 1M analytics events, 5K recordings, 1M flag requests, 100K exceptions, 1,500 survey responses, 10GB of logs, and 1M data warehouse rows. Usage stops at the limits, so an unexpected traffic spike cannot produce a surprise invoice. Pay-as-you-go keeps the same free tier and adds six projects and seven-year retention; credit add-ons appeared as USD 5, 25, and 20 packs in September 2026, and the vendor says 97 percent of companies stay on the free plan.

The open-source claim needs its footnote. The repository is MIT except the ee/ directory, which carries its own enterprise license, and GitHub lists the license as non-standard for that reason. Self-hosting is possible, and cloud regions in the US and EU cover data residency questions. The AI layer earns its place: PostHog AI queries the data in plain language across web, Slack, desktop, and MCP, and AI Observability plus Evaluations track LLM product usage, which matters for teams shipping AI features of their own.

Verdict

The default choice for product and marketing analytics when one tool should cover funnels, replay, flags, and surveys. The free tier is large enough to run a real product on, and the per-product billing limits keep costs bounded.

Similar Tools