PostHog Pricing 2026: Every Rate and Real Cost

Short answer: PostHog has no subscription fee on any plan. Paid is pay-as-you-go at a $0 base, so you pay only for usage above a monthly free allowance that renews every month: 1 million analytics events, 5,000 session recordings and 1 million feature flag requests, among others. A mid-size SaaS sending 5 million events, 20,000 recordings and 3 million flag requests pays about $365 a month. PostHog itself says 97% of its companies pay nothing at all.

Last updated September 2026. Every figure below was read at posthog.com/pricing and PostHog's public sales handbook on September 1, 2026. Where a number comes from somewhere else, it says so.

How PostHog pricing works

Most pricing articles about this category exist because the vendor hides something. This one exists for the opposite reason. PostHog publishes more of its commercial terms than any tool we have audited, including the per-unit rate for every product, the flat platform packages, the volume discount ladder, and the internal rule telling its own sales reps not to exceed it. The useful work here is not uncovering a hidden price. It is doing the arithmetic, because a rate card with fourteen separate meters is genuinely hard to turn into a number.

There are two plans and neither has a base fee.

Two details there are worth pausing on. There are no per-seat charges anywhere, which makes PostHog structurally cheaper than seat-priced analytics once your team grows. And on the free plan, usage stops at the limits rather than generating a bill, so the plan cannot surprise you. On paid, you set your own billing limit for each product separately, which is a meaningfully different arrangement from a single account-wide cap.

The full PostHog rate card

Every product meters something different and carries its own monthly free allowance. The allowance resets each month whether or not you have a card on file, and rates step down as volume rises.

Web analytics is bundled into product analytics at no extra charge, and experiments are billed with feature flags on the same meter and the same allowance. That last point matters more than it looks: A/B testing on PostHog is free up to a million flag requests a month, in a market where split testing starts at $99 a month on Crazy Egg and requires a sales conversation at Optimizely, AB Tasty and Adobe Target.

One line stands out for the wrong reason. Inbox, which opens pull requests for issues it finds, costs $15.00 per pull request after the first three. That is by a wide margin the highest unit price on the card, and it is the one meter where a small change in usage moves the invoice noticeably.

How much does PostHog cost per month?

Here is the arithmetic worked through at three sizes using the published tiers, so you can see where the money actually goes.

Take the mid-size row apart to see the tiering at work. Of 5 million events, the first million are free, the next million cost $0.00005 each for $50, and the remaining 3 million fall into the next bracket at $0.0000343 for $102.90. Of 20,000 recordings, 5,000 are free, 10,000 cost $0.005 for $50, and the last 5,000 drop to $0.0035 for $17.50. Flags work the same way. Annualized, that company pays about $4,385.

You will find third-party pricing aggregators quoting figures like $54,443 a year as a typical mid-size PostHog bill, and one that currently prints the per-event price as "$0.00 to $0.00". Those numbers have no visible source and do not survive contact with the published tiers. Treat any PostHog cost estimate that does not show its arithmetic as guesswork.

The add-on that quietly multiplies your event cost

This is the closest thing to a hidden cost on the card, and it is hidden only in the sense that it sits in a separate table further down the page.

Base product analytics counts anonymous events. If you want identified events, meaning person profiles that track users across visits and devices and let you build cohorts on user properties, that is a paid add-on with its own meter. It is priced at $0.000198 per event at the first paid tier, against $0.00005 for a base event. That is very nearly four times the rate.

So enabling the feature that almost every marketing and lifecycle use case depends on takes the same traffic from $50 to $248, a five-fold increase. Group analytics and data pipelines are structured the same way, as separate add-on meters on top of the base event. None of that is concealed, but a buyer who costs PostHog off the headline event rate alone will be wrong by a multiple rather than a margin.

Why single sign-on costs $250 a month

Usage is not where a PostHog bill usually goes wrong. Platform packages are. They are flat monthly fees that sit on top of usage, not instead of it.

The consequence is worth stating plainly, because it inverts the usual shape of a SaaS bill. If your usage sits inside the free tier and your security team requires single sign-on, your analytics cost nothing and your access control costs $3,000 a year. The platform package becomes 100% of your PostHog invoice. Role-based access control, which many mid-size companies treat as basic hygiene rather than an enterprise luxury, is $24,000 a year.

On this specific point PostHog loses to a competitor we have been hard on elsewhere: Heap includes SSO on its free plan. Amplitude gates SSO behind a sales-quoted tier and Mixpanel behind Enterprise, so PostHog is not unusual in the category. It is just not the winner here, and the transparency argument does not rescue it.

To be fair on the other side, PostHog does publish an enterprise price at all, which Optimizely, Adobe Target, AB Tasty, FullStory and Heap do not. A $2,000 figure you can budget against beats a quote you have to schedule a call to receive.

Does PostHog offer annual commitment discounts?

Yes, and it prints the schedule, which is the genuinely remarkable part. Discount ladders are normally the most guarded document in enterprise software because they are the sales rep's leverage. PostHog puts its own in a public handbook, with the instruction that reps should not go beyond it and that deals which do will not count toward quota.

Two levers stack on top. Paying additional contract years upfront adds 2.5% per year, capped at 5%. Committing in writing to a signature date, from the person who will actually sign, adds a one-time 5%. Requesting payment terms beyond Net 30 works in reverse and subtracts 2.5% for every 15 days, so Net 60 costs you five points. At the published ceiling the stack reaches 50%.

Three pieces of fine print will cost you money if you assume the intuitive version:

• The volume tier is a prerequisite. You cannot claim the timing or forecasting discounts without first qualifying for the volume one, which starts at $25,000 of credit.
• The tier is set by credit value, not cash paid, meaning the pre-discount list value of the credits you receive.
• It is measured per contract year, not per term. A two-year deal carrying $250,000 of credit each year sits in the 35% band, not the 40% one.

There is also a genuine inconsistency between PostHog's own two pages, and the less generous version is the accurate one. The pricing page summarizes the second lever as a two-year commitment earning 3% and a three-year commitment earning 5%. The handbook is more precise: the discount is 2.5% per additional year paid upfront, and it states explicitly that length of commitment is not a lever PostHog uses, because it rewards the prepayment rather than the signature. A three-year contract invoiced annually earns nothing extra. Terms are capped at 36 months by policy.

Almost none of this reaches a normal buyer, and it would be dishonest to present a $25,000 floor as a widely available discount. There is a smaller route: 10% off credit purchases below $25,000, offered in-app to customers with at least three paid invoices averaging $280 or more. Nonprofits get 15% below $25,000.

Can I use PostHog completely free forever?

For most companies, yes. PostHog states that 97% of its companies use it for free, that the allowance renews monthly on every product, and that this is not a trial. The free plan needs no credit card, carries unlimited team members, and stops collecting at the limits rather than billing you.

The constraints that actually bite are the ones that are not about volume: one project instead of six, one year of data retention instead of seven, and community rather than email support. For a single product with a normal team, none of those are likely to be what forces an upgrade. Running several distinct products under one organization is the common reason people move to paid, and that is a projects limit rather than a usage limit.

Startups can go further. PostHog gives $50,000 in credits for 12 months to companies under two years old that have raised under $5 million, and Y Combinator companies get $50,000 a year, renewable while they have raised under $25 million, from any batch. Set that against the $4,385 annual bill in our mid-size example and a qualifying startup would need to be an order of magnitude larger before it paid anything.

Can I self-host PostHog for free?

Yes, under an MIT license with a free Docker Compose deployment, and you should read the disclaimer before treating that as a cost-saving plan. PostHog states that self-hosted deployments are officially unsupported, that it does not do tagged releases, and that it does not publish CVEs at all, because self-hosted instances have no versions and every change including security fixes ships continuously from master. There are no guarantees and no paid support plans for self-hosted instances.

That makes self-hosting a reasonable answer to a data-residency or compliance requirement and a poor answer to a pricing one. The teams most tempted by it are usually the ones whose cloud usage would sit inside the free tier anyway, so they would be paying an engineer to maintain infrastructure that PostHog would have run for $0. The engineering time is the real bill, and it is bigger than the invoice it replaces.

If a developer shortage is what is actually blocking you, the instrumentation work is more delegable than it used to be. Deciding which events to capture and wiring an SDK into an existing codebase is a well-specified task, and it is the kind of change teams increasingly hand to an autonomous coding agent that plans the work and opens the pull request rather than queueing it behind a sprint. That does not change PostHog's price, but it does change how expensive its main prerequisite is.

Is PostHog worth it?

On price, the question is close to unanswerable in PostHog's favor, because for most teams there is no price. The more useful question is whether the shape fits you.

Stay if you have engineering time, want analytics, session replay, feature flags and experiments on one bill instead of four contracts, and have usage anywhere near the free tier. Nothing else in this category is cheaper or more transparent, and switching would waste a quarter to save nothing.

Look elsewhere if one of three things is true. You have no developer time and need something a marketer can operate. You need SSO or RBAC and cannot justify $250 to $2,000 a month for access control on top of usage that may cost nothing. Or what you actually wanted was a tool that tells you what to change rather than one that measures what happened. PostHog will tell you that 41% of visitors abandon on step two. It will not write the better step two.

That last gap is the one Converto fills, and it is deliberately a different job rather than a replacement: paste a live URL and it audits the copy, layout, CTAs, page speed and AI-readability, then generates the rewritten variants ready to ship, with nothing to install. Plenty of teams run both. For the wider comparison, see our roundup of the best PostHog alternatives, the head-to-head PostHog alternative page, or, if analytics pricing generally is what you are auditing, how Heap prices the same job.