PostHog is a product analytics platform designed to combine product data with AI agents to automatically diagnose issues, fix bugs, and generate pull requests. According to the official description, PostHog automatically diagnoses issues, fixes bugs, and generates pull requests without user prompts.
Purpose: PostHog serves as the context layer for a product, ingesting and storing analytics data, errors, replays, and business data so that users and agents can query and act on it. It provides a single place to ingest, store, and query product and company data, stitching together analytics, replays, errors, and logs in real time to answer any question. It can also import third-party data from over 900 sources for a more complete view.
Key capabilities:
- Product analytics: built-in product analytics features.
- AI observability: provides AI Observability.
- No-code A/B testing: supports no-code A/B testing.
- Data warehouse: built-in context warehouse, including a data warehouse, 120+ sources/destinations, SQL editor + BI + data visualization, and user activity streams (CDP-lite).
- Agent context: provides product context for agents, allowing product data to be queried from the editor without switching browsers. Supports PostHog MCP and PostHog Desktop.
- Self-driving pull requests: Self-driving pulls signals come from error tracking (exceptions and stack traces grouped into issues), session replay (dead clicks, rapid returns, long pauses), support (user tickets and conversations), and external tools (Zendesk, Linear, GitHub issues).
- Inbox: clusters related findings into research reports, sorted by priority. Review suggested improvements and pull requests, then decide what to ship.
How to use:
- Free setup: the setup wizard installs PostHog for the user, with no credit card required.
- Terminal alternative: if you prefer the terminal, you can skip the browser.
- Install PostHog MCP or use PostHog Desktop.
- Choose a cloud region: United States (Virginia) or European Union (Frankfurt).
Use cases:
- Use full context when analyzing customer usage or guiding AI.
- Combine everything in the PostHog context warehouse so that users, agents, and dashboards can query it directly.
- Includes data from 120+ external sources (such as Stripe, Postgres, HubSpot), as well as data from other PoInsights from stHog tools (such as session replay and experiments).
- Transform "tell me what happened" into "what to fix next."
Limitations:
- The demos mentioned in the official text are illustrative demos, and all findings are sample data.
- On pricing: all paid products are usage-based, with a generous monthly free tier, and 97% of customers use PostHog for free. The goal is to match the cheapest option at scale. No sales call required. Specific pricing examples are not provided for now.
- The official text does not provide explicit feature limitations or technical limitations, so none are provided for now.
Other information:
- More than 500,000 teams use PostHog.
- Transparency support: you can read the company handbook, sales handbook, and company strategy.
- Fast releases: check the changelog.
- Technical support: support staff all have engineering backgrounds.
- PostHog is a web product and cannot be installed via CD. Floppy disks were once sent to customers, but that was a Rickroll.
