PostHog vs Amplitude: pricing, free tiers, and where Amplitude vs PostHog actually differ

Amplitude is the better analytics product. PostHog is the better platform deal. That is the honest split, and which one wins depends on whether analysis depth or tool consolidation is the thing you are short of.

Amplitude has spent longer on the specific problem of behavioral analysis and it shows in the depth of the reporting: behavioral cohorts, advanced exploration, funnel and retention analysis that an analyst can push much further before hitting a wall. It is also the easier of the two for a non-engineer to work in unaided. PostHog is unmistakably built by engineers for engineers. Its analytics are perfectly capable and improving quickly, but a product manager without technical support will get value from Amplitude faster.

Where PostHog wins decisively is scope and price transparency. Feature flags, experiments, error tracking, a data warehouse, pipelines and logs arrive as products on the same bill, several of which would otherwise be separate vendors. For a team currently paying Amplitude plus LaunchDarkly plus Sentry, PostHog can collapse three invoices into one and publish what each part costs. The reviews reflect a preference split rather than a quality gap, and the Reddit threads that come up on this search mostly divide the same way: engineers pick PostHog, product managers pick Amplitude.

Yes, and it is the most complete free plan in product analytics. Amplitude's free tier gives you 2 million events a month forever with no credit card, 10,000 monthly session replays, unlimited seats, and every product line switched on: analytics, session replay, experimentation, guides and surveys, Amplitude AI, and 500 active AI Visibility prompts.

The limits are on depth rather than access. Experiments, guides and surveys are described as limited on Free, behavioral cohorts start at 20 on the Plus plan, heatmaps and alerts begin on Plus, and two year data retention starts on Plus. SSO and project permissions do not appear until Growth, which is custom quoted, so a company with a single sign-on requirement cannot stay in self-serve at any price it can look up. That last point is worth checking early, because it is the gate that most often forces a sales conversation that the buyer did not plan for.

One more thing worth knowing: if your company has raised under $10 million and has fewer than 20 employees, Amplitude's startup program gives you a free year on the Growth plan, which is otherwise entirely quote-based. That is a materially better offer than the free tier and it is easy to miss.

It depends on how many jobs you are trying to do for free. For analytics alone, Amplitude wins: 2 million events against PostHog's 1 million, plus 10,000 session replays against PostHog's 5,000 web recordings, plus two year retention once you move to Plus against PostHog's one year on free.

For a team trying to run several tools at zero cost, PostHog wins comfortably. Its free allowances apply to roughly eleven products at once: 1 million analytics events, 5,000 web and 2,500 mobile session recordings, 1 million feature flag requests, 1,500 survey responses, 100,000 error tracking exceptions, 1 million data warehouse rows, 10 GB of logs and 500 AI credits, every month. A small startup can genuinely run analytics, flagging, error tracking and surveys on PostHog without paying anyone.

The limit that ends each free plan is different, which is the practical thing to plan around. PostHog free accounts stop being viable because of the single project and one year retention cap, not the event volume. Amplitude free accounts stop being viable because someone needs SSO, deeper experimentation, or more than 20 behavioral cohorts. Work out which of those you will hit first and you will know which free tier is really free for you.

Amplitude is an analytics platform that bundles adjacent products into tiers. PostHog is a set of separately metered products that happens to include analytics. Everything else follows from that.

On Amplitude you pick a plan and everything in it is switched on, billed against one event meter. On PostHog there are no tiers at all. You add a payment method and pay for what each product consumes, with its own free allowance and its own sliding rate. That makes PostHog cheaper for teams who use a lot of one thing and nothing of another, and it makes Amplitude simpler for teams who want a single forecastable number.

There is a difference in origin that still shapes both products. PostHog is open source under an MIT license and grew out of a developer tool culture, which is why it ships error tracking, logs and a data warehouse. Amplitude grew out of the analytics needs of consumer product teams, which is why its cohort and exploration tooling is deeper and why it has added guides, surveys and now AI Visibility around the edges. Both are event-based, both require you to instrument your product, and neither can tell you why a number moved, because that answer is in tickets, reviews and survey replies rather than in the event stream.

These three come up as a set constantly, and they separate on one question: what is the scarce resource on your team?

If engineering time is scarce and money is not, shortlist Amplitude. It has the deepest analysis, the most complete free tier, and the shortest path from a non-technical question to an answer. If money is scarce and engineering appetite is not, shortlist PostHog. It is the cheapest of the three at volume, it publishes every rate so you can prove that before committing, and it removes separate flagging and error tracking vendors. If forecastability is the scarce thing, meaning finance wants one number they can defend, shortlist Mixpanel, which is the only one of the three with a flat published rate of $0.28 per 1,000 events that does not change with volume band or event type.

A practical tiebreak: check the identified event question on PostHog and the SSO question on Amplitude before you go further, because those two rows move more real quotes than any feature comparison does. The wider field, including the guidance-first tools like Pendo and Userpilot, is covered in our product analytics tools roundup, and each pairing has its own head to head worth reading before you commit budget.

You can, and unlike most pairings in this category it sometimes makes sense, because the two overlap less than their marketing suggests.

The arrangement that works is PostHog for the engineering products and Amplitude for the analysis: feature flags, error tracking and logs on PostHog, where they are cheap and well built, and behavioral analytics on Amplitude, where the free tier covers 2 million events and the reporting is deeper. That gives a product manager the interface they are fast in while the engineering team gets flagging and error tracking without a separate vendor. The cost of the arrangement is two instrumentation paths that will drift apart unless someone owns keeping them aligned.

What does not make sense is running both analytics products in parallel indefinitely. You will pay two meters for the same funnels and end up with two numbers that disagree, which is worse than either number alone. If you are running both because a migration stalled, finish the migration. The pairing that adds genuinely new information is a behavioral tool of either kind alongside something that reads the customer voice, because that is the gap neither of these two attempts to fill.