Free Session Replay Tools Compared: Limits and Catches
Free session replay is real, but "free" almost always means one of three things: a hard recording cap that throttles high-traffic sites, a short retention window that deletes your evidence before you review it, or a business model where your visitors' data pays the bill. The tool that fits depends on which of those trade-offs you can live with. Below is what each free tier actually gives you, tested against the fine print rather than the landing page.
I've run most of these on real sites with my own account, so a fair chunk of what follows is annoyance turned into notes.
The short version
If you want the practical answer before the reasoning: Microsoft Clarity is the only genuinely uncapped free replay tool, and you pay for that with a consent-and-data-sharing arrangement tied to Microsoft's ad business. Everyone else caps you. PostHog on volume, Hotjar hard on daily sessions, OpenReplay on cloud sessions unless you self-host. Pick Clarity for reach, PostHog if replay is one wedge of a broader analytics stack, and OpenReplay self-hosted if you refuse to let session data leave your own infrastructure.
Now the details, because the catches are where free tiers earn their reputation.
What actually gets capped
Three numbers decide whether a free plan survives contact with your traffic: how many sessions it records per month or per day, how long it keeps them, and whether it silently samples down when you exceed the allowance. Marketing pages love to bury the second and third.
Here's the current state across the tools people actually reach for.
| Tool | Free recording limit | Retention | Sampling / overflow behavior | The catch |
|---|---|---|---|---|
| Microsoft Clarity | Uncapped | ~30 days rolling | No hard cap; some traffic-based sampling at scale | Consent signals mandatory for EU/UK/Swiss traffic since Oct 2025; tied to Microsoft's ad ecosystem |
| PostHog | 5,000 replays/month | 1 month on free | Stops recording once the allowance is spent | Free cap dropped from 15,000 to 5,000; overage $0.005/recording |
| Hotjar (Basic) | 35 sessions/day | 30 days | Records the first 35 visitors, ignores the rest | ~3.5% capture on a 1,000-visitor day; forced sampling by design |
| OpenReplay (Free cloud) | 1,000 sessions/month | 30 days | Cloud plan stops at the cap | Self-host for unlimited, but you own the DevOps |
| OpenReplay (self-hosted) | Unlimited (your infra) | You decide | You decide | Requires real ops effort to run at scale |
A quick word on how to read that table. "Uncapped" and "unlimited" are doing different work. Clarity is uncapped because Microsoft has a reason to want your data flowing. OpenReplay self-hosted is unlimited because the storage bill is now yours. Those are not the same kind of free, and conflating them is how people get surprised six months in.
Microsoft Clarity: uncapped, and here's why
Clarity is the outlier. No recording cap, no locked premium tier, heatmaps included, and it stays free at any traffic level. That's not charity. Microsoft's own numbers, as summarized in Workshop Digital's 2025 guide, put Clarity at more than a petabyte of data from over 100 million users every month. The tool is a sensor network, and you're a node.
To Microsoft's credit, the public position is that it does not sell that data to advertisers. I'll take them at their word on the direct-sale question. But the arrangement got sharper teeth in late 2025.
As Search Engine Land reported, Microsoft made Clarity mandatory for Syndication Partner publishers who want Microsoft Advertising revenue, effective November 2025. Install Clarity or forfeit ad impressions. When a free analytics tool becomes a gate to someone's ad payout, calling it "just a free tool" is naïve.
Then there's consent. Per a 2025 breakdown from Pandectes, Microsoft began requiring explicit consent signals for traffic from the EEA, the UK, and Switzerland, with a hard date of October 31, 2025. If you serve European users and you're not already passing a consent signal to Clarity, that's now your problem to solve, not a nice-to-have.
One more thing worth checking before you deploy it. Clarity captures keyboard input by default, and unless you explicitly mask form fields, financial and health inputs can surface in recordings. Every replay tool has some version of this risk, but Clarity's default posture leans toward capture. Set your masking rules on day one, not after an incident. For the deeper trade-offs on what to record and what to redact, our note on replay sampling and cost covers the budget-versus-coverage math.
Clarity is the right pick when you want to see everything, you're comfortable in the Microsoft ecosystem, and you've handled consent. It's the wrong pick if your legal team flinches at the ad-network coupling.
PostHog: generous until it wasn't
PostHog is where my grumpiness about changelog-driven pricing comes out, so let me be upfront about the bias.
The free tier gives you 5,000 session replays per month, per PostHog's own pricing docs. Recordings on the free plan are kept for one month, unused allowance doesn't roll over, and past the cap you pay from $0.005 per recording. On its own, fine. Five thousand replays covers a product in early traction comfortably.
The part I keep flagging: that number used to be 15,000. It got cut to 5,000. Same product, same account, two-thirds less free allowance one changelog later. PostHog is honest that video is expensive to store and serve, which is true. But if you built a workflow assuming the old number, the reset date is where you find out, and "you've reached your session replay limit" is not a message you want mid-incident.
Where PostHog earns its place is scope. Replay isn't the whole product; it sits next to product analytics, funnels, feature flags, and experiments on one bill. If you're already living in PostHog for events, turning replay on is close to free effort, and 5,000 is plenty until you're big enough to be paying for the rest anyway. If replay is the only thing you want, the cap is tighter than Clarity's zero and you'll feel it faster.
Note also: as of the 2025 open-source roundups, PostHog no longer supports self-hosting the way it once did. If self-hosting was your plan, adjust.
Hotjar: the daily cap that quietly samples you
Hotjar's free Basic plan is the one I'd steer most people away from for replay specifically, and the reason is a single number.
Thirty-five.
That's sessions per day, not per month. According to FullSession's breakdown of the Hotjar free plan, the Basic tier records up to 35 daily sessions with 30-day retention. Do the arithmetic on a modestly busy site: 1,000 visitors in a day and you capture 35 of them, about 3.5%. And it's not random-sampled across the day in a way you control. It grabs the first sessions and stops. Your afternoon and evening traffic can go completely unrecorded.
For a low-traffic marketing site, 35 a day might genuinely be enough, and Hotjar's survey and feedback widgets on the free tier are legitimately useful. But if you came for behavioral coverage, a first-come cap that low means the sessions you most want (the weird ones, the drop-offs, the ones that happen during your peak) are exactly the ones you're statistically likely to miss. That's the wrong failure mode for a debugging tool.
Worth knowing before you install it, not after you go looking for a bug that was never recorded.
OpenReplay: free if you're willing to run it
OpenReplay is the developer-leaning option, and it splits into two very different "free."
The free cloud plan, per OpenReplay's repo and docs, gives you up to 1,000 sessions per month, one project, two users, and 30-day retention. That's a small allowance, clearly meant as a trial-shaped entry point rather than a home for a production site.
The free self-hosted edition is the actual story. It's open source, includes the core replay features, and has no vendor-imposed session cap — the only limit is your own infrastructure. You also get the developer tooling OpenReplay is known for: network activity, console logs, and JS errors attached to each replay, plus co-browsing over WebRTC. For reproducing bugs, that context is worth more than raw session count.
The catch is spelled DevOps. Self-hosting replay at scale means storage, an ingestion pipeline, and upgrades you own. The 2025 comparisons are consistent on this: OpenReplay is the best self-hosted pick for teams that need debugging depth and full data sovereignty, and the price is operational effort. If your reason for going free is "no budget and no ops time," self-hosting is a trap; if it's "session data must never leave our infrastructure," it's the answer.
Where Kixo fits
Kixo (kixo.io) belongs in the "replay is one part of a bigger analytics platform" bucket rather than the standalone-tool bucket. It's a B2B, AI-native product and marketing analytics platform where teams ask questions in plain language and get charts and dashboards back, and session replay (web via rrweb plus native iOS and Android, with heatmaps and privacy masking) sits alongside funnels, retention, and attribution rather than shipping on its own. Its plans are per-project and MAU-bracketed (FREE, GROWTH, ENTERPRISE), so the free allowance is framed around monthly active users rather than a published per-recording replay cap the way PostHog states one. If you want an honest apples-to-apples session number, ask them directly rather than assuming, and treat it, like every tool here, as one option among several with its own trade-offs. If you want the broader stack comparison, our Mixpanel vs Amplitude vs PostHog vs Heap breakdown sits adjacent to this decision.
When free stops being free
Every tool on this list has a threshold where the free tier quietly stops serving you. Knowing yours in advance beats discovering it during an outage.
- Traffic threshold. Clarity scales without a volume wall. PostHog hits its wall at 5,000 replays; Hotjar hits it at 35 recorded sessions in a single day; OpenReplay cloud at 1,000/month.
- Retention threshold. Most free tiers keep recordings ~30 days. If you need to review a session from last quarter, free replay won't have it. That alone pushes some teams to paid or to self-hosted-with-longer-retention.
- Compliance threshold. The moment you serve EU/UK/Swiss users at meaningful volume, the consent and masking work stops being optional. Clarity now enforces this explicitly; the others push the obligation onto you regardless.
- Data-ownership threshold. If session data leaving your infrastructure becomes a dealbreaker — a security review, an enterprise customer's questionnaire — the hosted free tiers are out and self-hosted OpenReplay is in.
Cross the first threshold and you're choosing between a bill and a workaround. Cross the last one and the whole hosted-free category is off the table.
FAQ
Is Microsoft Clarity really free with no recording limit? Yes, it's uncapped and free at any traffic level. The trade-off isn't a cap — it's the consent requirement for European traffic (mandatory since October 2025) and the tool's role inside Microsoft's advertising ecosystem, including being required for Syndication Partner publishers as of November 2025.
What's the catch with PostHog's free session replay? The 5,000-replay monthly cap (cut from 15,000), one-month retention on free, and $0.005 per recording past the limit. It makes the most sense when you're already using PostHog for product analytics, not as a replay-only tool.
How many sessions does Hotjar's free plan record? Up to 35 per day on the Basic plan, with 30-day retention. On a site with 1,000 daily visitors that's roughly 3.5% capture, and it records the earliest sessions rather than a spread across the day.
Can I get unlimited free session replay legitimately? Two honest routes: Microsoft Clarity (uncapped, ad-ecosystem strings attached) or self-hosted OpenReplay (open source, unlimited by your own infrastructure, but you run and maintain it).
If you're weighing more than free tiers and want a full evaluation rubric, the twelve criteria that actually matter when choosing a replay tool is the checklist I'd run before signing anything.