What is PostHog?
PostHog is a product analytics and data platform that ingests and stores analytics, errors, session replays, and business data so teams and their AI agents can query and act on it from a single place. PostHog differentiates itself by stitching that data together on the fly and pairing it with AI that can diagnose problems, fix bugs, and generate pull requests without being prompted.
Key Features
- Product analytics: PostHog captures and stores product analytics data alongside errors, replays, and business data, so teams can analyze funnels and user behavior without stitching together separate tools.
- Session replay and Replay Vision: Session replays record user sessions, and Replay Vision surfaces issues found in those replays, such as customers retrying the same declined card.
- Error tracking and logs: PostHog ingests errors and logs next to analytics data, allowing problems like mobile sign-in failures to be traced and clustered with related findings.
- Feature flags and experiments: PostHog includes feature flags, experiments, and no-code A/B testing for launching and measuring changes.
- Surveys and web analytics: Surveys and web analytics are built into the same platform as the rest of the product data.
- Data warehouse and CDP: A warehouse and CDP layer let teams pipe in third-party data from 900+ sources for a more complete picture, alongside workflows, endpoints, tracing, heatmaps, and AI observability.
- PostHog AI and Inbox: PostHog AI answers questions across product data, while Inbox clusters related findings into researched reports ranked by priority, with proposed fixes and draft pull requests.
- MCP and editor integration: The PostHog MCP lets agents query product data from an editor instead of switching to a browser, and PostHog Desktop offers an alternative client. Setup is handled by a wizard, installable via
npx -y @posthog/wizard@latestornpx @posthog/wizard mcp addfor MCP.
Use Cases
- Debugging conversion drops: A team investigating falling paid conversions can have PostHog AI compare Google Ads campaign traffic, analyze the signup-to-payment funnel, check Stripe payments and declines, and review recent GitHub checkout changes to locate the cause.
- Prioritized issue triage: Engineering teams can use Inbox to receive clustered findings, such as a high-priority sign-in error with a pull request ready to review, ranked by priority.
- Agent-driven development: Developers can connect the PostHog MCP so agents query product data from their editor and handle everything from one-off analytics to launching new features without a separate UI.
- Unified product and business analysis: Teams can combine first-party analytics, replays, errors, and logs with third-party data from 900+ sources to answer questions across the whole product and company data set.
- Experimentation and rollout: Product teams can run feature flags, experiments, and no-code A/B tests to ship and measure changes.
Who It Is For
PostHog is aimed at product teams, engineers, and data teams that need analytics, session replays, error tracking, feature flags, and business data in one place rather than spread across separate tools. It also targets developers and teams working with AI agents, since PostHog provides the product context those agents need to investigate issues and propose fixes directly from an editor or inbox.
Pricing
PostHog offers a free start with no credit card required, and the source states that 97% of users pay $0. Specific plan prices are not listed in the source material.
Frequently Asked Questions
What is PostHog used for?
PostHog is used to ingest and store product analytics, session replays, errors, logs, and business data in one place, then query and act on it. Teams use PostHog to analyze funnels, investigate conversion drops, track errors, run feature flags and experiments, and let AI agents diagnose problems and generate pull requests.
Is PostHog free?
PostHog can be set up for free with no credit card required, and the source states that 97% of users pay $0. Specific paid plan details are not provided in the source material.
How does PostHog work with AI agents?
PostHog acts as a context layer for agents: it stores analytics, errors, replays, and business data that agents can query and act on. Developers can install the PostHog MCP to query product data from their editor, and PostHog AI plus Inbox can investigate issues and produce researched reports with draft pull requests.
What data sources can PostHog connect to?
PostHog ingests its own analytics, replays, errors, and logs, and can pipe in third-party data from 900+ sources through its warehouse and CDP capabilities. The source also references integrations with tools such as Google Ads, Stripe, and GitHub in its investigation workflow.
Who should use PostHog?
PostHog suits product teams, engineers, and data teams that want analytics, session replay, error tracking, feature flags, experiments, and business data unified in one platform. It is also relevant for developers who want AI agents to work from real product context inside their editor.
