OpenReplay
Self-hosted session replay where the data never leaves your cloud
OpenReplay is an open source session replay and product analytics platform that captures user sessions alongside network activity, console output, JavaScript errors, and application state, and can be self-hosted on your own infrastructure for free or run as a managed dedicated instance from $199 a month; because it is deployed into your own cloud account in the self-hosted mode, no user data is sent to a third-party processor at all.
Overview
OpenReplay is what happens when a session replay tool is designed around the objection that stops most engineering teams from buying one: sending a full recording of every user's screen, network traffic, and application state to a third-party SaaS company. OpenReplay's answer is to hand you the software. You deploy it into your own AWS, GCP, Azure, DigitalOcean, Scaleway, or OVHcloud account, or any Kubernetes cluster, and the recordings live in your infrastructure, under your retention policy, in your region, with no data processing agreement required because there is no processor.
The company began in 2018 as Asayer, went through Y Combinator's Summer 2019 batch, and pivoted from closed-source SaaS to open source in 2020, which turned out to be the decision that defined it. It raised a $4.7M seed round in 2022 led by Runa Capital with Expa, 468 Capital, and Rheingau Founders participating, bringing total funding above $6M, which is a fraction of what its competitors have raised and consistent with a company whose distribution comes from a GitHub repository rather than a sales team. That repository sits at roughly 12,000 stars, and releases continue on a regular cadence.
Functionally it is closer to LogRocket than to Hotjar. The tracker is around 26KB and captures network requests, console logs, JavaScript errors, over 40 performance metrics, and application state from Redux, VueX, MobX, and similar stores, plus GraphQL and REST logging. Around the replay sit product analytics with funnels and journeys, heatmaps, Assist for live WebRTC co-browsing with a user in trouble, and Spot, a Chrome extension that lets anyone record a bug with full technical context and hand it to a developer as a link.
The pricing model is the second thing that makes it unusual. The managed Dedicated tier is billed by instance rather than by session: $199 a month, or $0.276 an hour, for a fully managed instance with no limits on users, recordings, or retention. That is a fundamentally different cost curve from every session-metered competitor, because your bill stops tracking your traffic. The obvious cost is elsewhere: self-hosting means you operate a Kubernetes deployment with real storage requirements, and that is engineering time nobody is billing you for but somebody is spending.
Best for
Engineering teams who want LogRocket-class session replay and frontend debugging but cannot or will not send user session data to a third-party processor, and who have the infrastructure competence to run a Kubernetes deployment or the budget for a $199 a month managed dedicated instance.
Not the right fit for
- Teams with no infrastructure capability who want the free version; self-hosting is a real Kubernetes deployment with storage, scaling, and upgrade responsibilities, and treating the open source edition as free software with no operational cost is how it goes badly.
- Non-technical marketing teams optimising landing pages; the product is built for engineers, and Microsoft Clarity or Mouseflow will do that job faster and with less to learn.
- Companies wanting a full product analytics platform; the analytics are a useful companion to the replay rather than a substitute for Mixpanel, Amplitude, or PostHog, with no behavioural cohort modelling or experimentation.
- Mobile-first products; OpenReplay is a web-centric platform and its mobile story is far thinner than its web one, unlike Mixpanel and Amplitude which capture iOS, Android, and React Native replay directly.
- Buyers who need a large vendor behind the contract; OpenReplay has raised roughly $6M in total and is small, so a procurement process that weighs vendor size will not favour it, even though the open source code is a genuine hedge against the company disappearing.
How it works
- 1
You choose a deployment. Self-hosted means installing OpenReplay into your own cloud account or Kubernetes cluster, where the recordings, the database, and the object storage all live in your infrastructure. Dedicated means OpenReplay operates a managed instance for you across a wide selection of regions, with a bring-your-own-cloud option for teams who want managed operations but their own account.
- 2
You add the tracker to your web application, roughly 26KB, and it begins recording sessions as a serialised DOM stream rather than as video. Alongside the visual recording it captures network requests and responses, console output, JavaScript errors with stack traces, over 40 performance metrics, GraphQL and REST calls, and state from Redux, VueX, MobX, and comparable stores, which is the layer that makes a recording useful to a developer rather than only to a designer.
- 3
Sanitisation is configured in the tracker, so sensitive fields are obscured before anything is transmitted. Because the destination in self-hosted mode is your own storage, the privacy question changes shape entirely: you are deciding what your own systems retain rather than what you are comfortable sending to a vendor.
- 4
In the console, omni-search filters sessions by user actions and technical events together, so you can find every session where a specific API returned an error inside a particular funnel step. Analytics adds funnels, journeys, trends, and heatmaps over the same data. Assist lets a support engineer watch a live session and join a WebRTC call with remote control and annotation, and Spot lets anyone on the team capture a bug from their browser with the technical context attached. An MCP server exposes the data to AI assistants.
Feature breakdown
23 features in 4 modulesSession replay and developer tools
The core capability, with the technical context an engineer needs.- DOM-based session replay
- Sessions are captured as a serialised DOM and interaction stream rather than as video, giving faithful, searchable, lightweight playback with a tracker of around 26KB so page performance is not the price of visibility.
- Network request capture
- Requests and responses including GraphQL and REST calls, with timing and status, recorded in sync with the replay so a failure can be attributed to a specific call rather than guessed at.
- Console and error capture
- Console output at every level plus JavaScript exceptions with stack traces, tied to the moment they occurred in the session.
- Application state inspection
- Redux, VueX, MobX, and comparable stores are recorded action by action, so a developer can step through the exact state transitions that produced a broken render in a real user's browser.
- Over 40 performance metrics
- Page timings, resource loading, memory, and CPU behaviour captured per session, turning vague reports of slowness into a specific, attributable measurement.
- Omni-search across behaviour and technical events
- One search surface filtering by user actions and technical events together, which is how you find every session where a particular endpoint failed during checkout rather than searching those two things separately.
Product analytics
Enough analysis to act on the recordings without buying a second tool.- Conversion funnels
- Multi-step funnels with the ability to open the sessions behind any drop-off, which is the natural advantage of analytics built on top of complete recordings.
- User journeys
- Aggregate path analysis showing the routes users actually take, including the ones the design did not anticipate.
- Trends and metrics dashboards
- Custom metrics and dashboards over captured behaviour and technical signals, so error rates and conversion rates can sit on the same board.
- Heatmaps
- Click and interaction heatmaps generated from the same session data, without a second tag or a separate sampling population.
- Automatic issue detection
- Surfacing of the issues affecting conversion and revenue, including frustration signals and errors correlated with drop-off, so triage starts from a list rather than from a hunch.
- Web analytics
- Basic traffic reporting alongside the product analytics, which for a small team removes the need for a separate lightweight analytics tag on the same property.
Support and collaboration tools
The parts that turn replay into a way of helping a specific user right now.- Assist live co-browsing
- Watch a user's session live and join a WebRTC call with them, with remote control and on-screen annotation, so a support engineer can guide someone through a problem rather than describing it over email.
- Spot bug capture extension
- A Chrome extension that lets anyone record a bug from their own browser with the technical context attached and share it as a link, which raises the quality of every internally filed bug report.
- MCP server
- A Model Context Protocol server exposing session and analytics data to AI assistants, so an agent can query what happened in production rather than being told about it.
- Issue tracker and support integrations
- Session links flow into the tools where work already happens, so a recording is attached at the point a ticket is created rather than hunted for afterwards.
Deployment and data control
The reason to choose OpenReplay over a hosted competitor.- Self-hosting on your own cloud
- Deployable to AWS, GCP, Azure, DigitalOcean, Scaleway, OVHcloud, or any Kubernetes cluster, with documented installation guides per platform. In this mode no session data reaches OpenReplay at all.
- No third-party processor
- Self-hosting removes the vendor from the data path entirely, which is what makes lengthy compliance review unnecessary rather than merely faster: there is no processor to assess and no cross-border transfer to justify.
- Managed dedicated instances
- A fully managed instance from $199 a month, billed hourly at $0.276, with no limits on users, recordings, or retention, available across roughly 50 regions.
- Bring your own cloud
- Managed operation of an instance running inside your own cloud account, which splits the difference between operational burden and data control.
- Custom retention policies
- You set how long recordings are kept, rather than accepting a vendor's plan-tier default. On self-hosted deployments this is entirely a storage decision you own.
- Open source code
- The full platform is published on GitHub under a monorepo using several licenses, with roughly 12,000 stars and continuing releases, so the code is auditable and the product survives the company in a way a SaaS subscription does not.
- Tracker-level sanitisation
- Sensitive content is obscured in the tracker before transmission, with per-element controls, so sanitised data is never written to storage rather than being hidden after arrival.
Use cases
4 documentedFintech or healthtech engineering team
The team wants session replay, but legal will not approve sending recordings of authenticated user screens and network traffic to another company.
OpenReplay is deployed into the company's own AWS account. Recordings never leave the perimeter, no processor agreement is required, and the compliance conversation that killed the LogRocket purchase never has to happen.
High-traffic product tired of session-based pricing
Recording every session on a competitor's meter would cost several thousand dollars a month, so the team samples aggressively and misses the sessions they actually needed.
A dedicated managed instance at $199 a month with no limits on recordings or retention decouples cost from traffic entirely, and the team records everything again.
Support team handling complex product issues
Support cannot see what the customer sees and every ticket becomes a slow exchange of screenshots and guesses.
Assist lets an engineer watch the live session and join a WebRTC call with remote control and annotation, resolving in one conversation what previously took four emails and two days.
Engineering team drowning in low-quality bug reports
Internal bug reports say it broke with no reproduction steps, no environment detail, and no idea what the network was doing.
Spot lets anyone capture the bug from their own browser with console, network, and state context attached, and the resulting ticket is actionable on arrival.
Pricing
from $0 self-hosted, or $199 per month for a managed Dedicated instanceFree and open source when self-hosted, with managed hosting billed per instance and per hour rather than per session. Three options: self-hosted open source, a managed Dedicated instance, and Enterprise contracts for self-hosted deployments at scale.
| Plan | Price | Includes |
|---|---|---|
| Open Source (self-hosted) | $0 software cost, plus your own infrastructure |
The software is free; running it is not. Budget engineering time for deployment, storage, scaling, and upgrades. |
| Dedicated (managed cloud) | From $199 per month, billed hourly at $0.276 per hour |
Instance-based rather than session-based billing, which is the single most important commercial difference from every competitor in this category. |
| Enterprise (self-hosted at scale) | Quoted on multi-year contracts annual or multi-year |
|
Add-ons
- Bring your own cloud (Included in Dedicated arrangements): OpenReplay operates the instance while it runs inside your cloud account, so you get managed operations without the data leaving your perimeter.
- Your own infrastructure cost (Whatever your cloud provider charges): Not an OpenReplay charge, but the real cost of the free tier. Session recordings are storage-heavy, and object storage plus compute for a busy site is not trivial.
Billing notes
- The meter is instance time, not sessions. At $0.276 an hour, or roughly $199 a month, cost does not rise as your traffic does, which inverts the economics of every session-priced competitor.
- Self-hosting has no software licence cost at all, but the infrastructure cost is real and grows with recording volume and retention, because full session recordings consume meaningful object storage.
- Retention is your decision rather than a plan attribute. Self-hosted deployments retain what your storage budget allows, and Dedicated advertises no retention limit, which is unusual.
- The 7 day trial on Dedicated requires no credit card, and the open source edition needs no trial at all, so evaluation costs nothing but time.
- Enterprise pricing is on multi-year contracts with volume pricing, aimed at self-hosted deployments handling terabytes a day rather than at small businesses.
- The monorepo uses several licenses rather than a single permissive one, so if the licence terms matter to your legal team, read the LICENSE file before assuming standard open source rights across every component.
Value assessment: OpenReplay has the most favourable cost curve in this category and the least favourable setup cost, and which dominates depends entirely on your team. At 10,000 monthly users generating maybe 25,000 sessions, a competitor costs roughly $176 a month and OpenReplay Dedicated costs $199, so there is no saving and the reason to choose it is data control rather than money. At 100,000 monthly users generating a quarter of a million sessions, the competitor's bill runs into four figures monthly while OpenReplay Dedicated stays at $199, and the saving becomes the dominant argument. Self-hosted, the software cost is zero at any scale and you pay your cloud provider instead, which for a team already running Kubernetes is a small marginal cost and for a team that is not is a project. The honest summary is that OpenReplay is dramatically cheaper at scale, roughly break-even at small scale, and worth it at any scale if the alternative is not having replay at all because compliance said no.
Strengths & limitations
Strengths
- Self-hosting removes the third-party processor from the data path entirely, which converts session replay from a compliance argument into an infrastructure decision.
- Instance-based pricing at $199 a month with no limits on users, recordings, or retention decouples cost from traffic, which no session-metered competitor can match at scale.
- Technical depth genuinely comparable to LogRocket: network and GraphQL capture, console output, stack traces, over 40 performance metrics, and Redux, VueX, and MobX state inspection.
- A tracker of around 26KB, which is small enough that adding replay does not become a page performance conversation.
- Open source on GitHub with roughly 12,000 stars and continuing releases, so the code is auditable and the product outlives the company if the company does not.
- Assist provides live WebRTC co-browsing with remote control and annotation, a capability most replay tools do not have at all and which changes what a support team can do.
- Spot lets non-engineers file bug reports with full technical context attached, which quietly improves the quality of every ticket in the organisation.
- SOC 2 Type II compliance on the managed side, plus deployment options spanning six major clouds and any Kubernetes cluster.
Limitations
- Self-hosting is a real operational commitment: Kubernetes, object storage, scaling, and upgrades are yours, and session recordings consume storage faster than teams expect.
- Web-centric. Mobile replay is not the strength here, and teams needing iOS, Android, and React Native capture will find Mixpanel or Amplitude better served.
- Product analytics are a companion rather than a platform: no behavioural cohort modelling, no experimentation, no feature flags, and less analytical flexibility than a dedicated event tool.
- The company is small, with roughly $6M raised in total against competitors who have raised tens or hundreds of millions, which matters to procurement even though the open source code hedges the risk.
- The monorepo carries several licenses rather than one, so open source rights vary by component and should be checked rather than assumed.
- At small scale the managed tier is not cheaper than session-priced competitors, since $199 a month buys around the same as $176 a month elsewhere, so the argument at that size is data control rather than economics.
- Documentation and community are good but smaller than the well-funded alternatives, so obscure problems have fewer existing answers to find.
Head-to-head comparisons
4 alternativesOpenReplay vs LogRocket
from $0 (Free, 1,000 sessions per month), Core from around $176 per month at 25,000 sessionsThe direct comparison, and a genuine one. LogRocket is more polished, has deeper framework state integration and the Galileo machine learning prioritisation layer, and requires no operations. OpenReplay gives you the same category of technical depth with self-hosting, open source code, and a flat instance price. Choose LogRocket if you want the best experience and can accept US-hosted data on a session meter; choose OpenReplay if data residency, cost at high traffic, or code auditability outweighs polish.
Full OpenReplay vs LogRocket comparisonOpenReplay vs Microsoft Clarity
from $0Opposite ends of the same category. Clarity is free, requires no operations, and gives marketers heatmaps and frustration signals with 30 day retention and no engineering detail. OpenReplay gives engineers network, console, and state capture with retention you control, at the cost of either $199 a month or an infrastructure project. Run Clarity on the marketing site and OpenReplay in the application; they are not competing for the same job.
Full OpenReplay vs Microsoft Clarity comparisonOpenReplay vs Mouseflow
from $0 (Free, 500 sessions per month), then $25 per month (Essential)Mouseflow is a conversion optimisation tool for marketers with form analytics, friction scoring, and heatmaps from $25 a month, priced on recorded sessions. OpenReplay is a developer tool with network and state inspection, priced per instance or free self-hosted. If your question is which form field loses people, Mouseflow is cheaper and better shaped. If it is why the API call fails for eight percent of users, only OpenReplay answers it, and it answers it without sending anything to a vendor.
Full OpenReplay vs Mouseflow comparisonOpenReplay vs Matomo
from Free self-hosted (Community edition), or 29 euros per month for Matomo Cloud at 50,000 hitsBoth are open source, self-hostable, and chosen primarily by teams who do not want behavioural data leaving their control. Matomo is web analytics first with session recording and heatmaps attached, aimed at marketers and GDPR-conscious organisations. OpenReplay is session replay first with developer tooling, aimed at engineers. A privacy-focused company frequently runs both: Matomo for traffic and OpenReplay for the application.
Full OpenReplay vs Matomo comparisonImplementation & onboarding
- Setup time
- Fifteen minutes on a managed Dedicated instance: sign up, add the tracker, and sessions appear. Self-hosting is a half day to a few days depending on your Kubernetes fluency and cloud, plus ongoing operational attention for storage growth and upgrades.
- Learning curve
- Low for developers, who will recognise the devtools interface immediately. Higher for the self-hosted deployment itself, which is a normal infrastructure exercise rather than a product to learn. Non-technical staff can watch replays but will not use the network or state panels.
- Onboarding
- Entirely self-serve on both paths: the open source edition installs from documented per-cloud guides, and the Dedicated tier offers a 7 day trial with no credit card. Enterprise contracts add dedicated support and implementation help for large self-hosted deployments.
- Migration notes
- Session recordings do not port between vendors, so migration from LogRocket, FullStory, or the discontinued Smartlook means running in parallel for a period and then cutting over. The bigger migration decision is architectural rather than data-related: moving to self-hosted OpenReplay means adopting an infrastructure responsibility you did not previously have. Note that Smartlook, a former alternative in this space, reached end of sale in May 2026 and is being decommissioned, so teams looking for a replacement have a real deadline.
Platform, API & security
- Platforms
- Web via a tracker of roughly 26KBReact, Angular, Vue, and Svelte applicationsRedux, VueX, and MobX state integrationsSelf-hosted on AWS, GCP, Azure, DigitalOcean, Scaleway, OVHcloud, or any Kubernetes clusterManaged dedicated instances across roughly 50 regionsSpot Chrome extension for bug capture
- API
- REST API for sessions and analytics data, an MCP server exposing the platform to AI assistants, and full source access on GitHub for anything the API does not cover.
- Compliance
- SOC 2 Type IIGDPR, with self-hosting removing the third-party processor entirelyCCPASelf-hosted deployments inherit whatever compliance posture your own infrastructure holds
- Data residency
- Complete control in self-hosted mode: the data lives wherever your cloud account is, in whatever region you choose, subject to your retention policy. Managed Dedicated instances are available across roughly 50 regions, with a bring-your-own-cloud option for managed operation inside your own account.
- SSO
- Single sign-on available on the Dedicated tier; SAML and SCIM provisioning on Enterprise contracts.
- Security notes
- Sanitisation is applied in the tracker before transmission, so obscured content is never written to storage. Self-hosted deployments mean no session data reaches OpenReplay at any point. Audit logging is available on Enterprise contracts, and the tracking and platform code are published on GitHub for audit.
Support & resources
- Channels
- Community Slack and GitHub issues for open source usersEmail support on the Dedicated tierDedicated support on Enterprise contracts
- Documentation
- Detailed technical documentation covering tracker installation, per-cloud self-hosting guides, sanitisation configuration, framework plugins, Assist, Spot, and the MCP server.
- Community
- An open source community around a GitHub repository with roughly 12,000 stars and 700-plus forks, an active Slack, and public release notes; smaller than the well-funded competitors but engaged and technically capable.
Company
- Founded
- 2018
- Headquarters
- Distributed, incorporated in the United States as Asayer, Inc.
- Ownership
- Venture-backed, privately held, with the platform published as open source
- Founders
- Mehdi Osman
- Employees
- Not disclosed; a small team consistent with roughly $6M raised
- Funding
- More than $6M raised, including a $4.7M seed round in 2022 led by Runa Capital with Expa, 468 Capital, Rheingau Founders, and Y Combinator among the backers.
Funding history
| Round | Amount | Year | Notes |
|---|---|---|---|
| Y Combinator | Standard YC investment | 2019 | Summer 2019 batch, then operating as Asayer with a closed-source SaaS product. |
| Seed | $4.7M | 2022 | Led by Runa Capital with Expa, 468 Capital, Rheingau Founders, and the co-founders of Tekion participating, two years after the pivot to open source. |
Timeline
- 2018Founded by Mehdi Osman as Asayer, Inc., building a hosted session replay product for developers.
- 2019Goes through Y Combinator's Summer 2019 batch.
- 2020Pivots from closed-source SaaS to open source, publishing the platform on GitHub and reframing the product around self-hosting and data ownership. This is the decision the company is now built on.
- 2022Raises a $4.7M seed round led by Runa Capital, with the open source repository serving as the distribution engine rather than a sales team.
- 2024Adds Spot, a Chrome extension for capturing bugs with full technical context, and expands the analytics layer with funnels, journeys, and heatmaps.
- 2025Ships an MCP server exposing session and analytics data to AI assistants, and moves managed hosting to instance-based pricing with no limits on recordings or retention.
- 2026Continues regular releases with the GitHub repository at roughly 12,000 stars, as Smartlook's shutdown and Hotjar's absorption into Contentsquare push more teams to look for independent replay options.
Integrations
- Sentry, Bugsnag, Rollbar, and Datadog
- Jira, GitHub, and Linear
- Slack and Microsoft Teams
- Zendesk and Intercom
- Segment
- Elasticsearch and CloudWatch for log correlation
- Redux, VueX, MobX, GraphQL, and REST instrumentation
- MCP server for AI assistants
Frequently asked questions
10 questionsWhat is OpenReplay?
OpenReplay is an open source session replay and product analytics platform. It records what users do in your web application alongside network requests, console output, JavaScript errors, over 40 performance metrics, and application state from Redux, VueX, or MobX. It can be self-hosted on your own infrastructure for free or run as a managed dedicated instance, and it includes live co-browsing, a bug capture extension, and an MCP server.
How much does OpenReplay cost?
The self-hosted open source edition is free; you pay only for your own cloud infrastructure. The managed Dedicated tier starts at $199 a month, billed hourly at $0.276, for a fully managed instance with no limits on users, recordings, or retention, and comes with a 7 day trial requiring no credit card. Enterprise self-hosted contracts at scale are quoted on multi-year terms.
Why is instance-based pricing a big deal?
Because every other tool in this category meters sessions, so your bill rises with your traffic and the natural response is to sample and miss the sessions you needed. OpenReplay's managed tier charges for the instance, not the recordings, so a site with 20,000 sessions and a site with 2 million sessions pay the same $199. At small scale that is roughly break-even against competitors; at high traffic it is the difference between $199 and several thousand dollars a month.
Is self-hosting actually free?
The software is, the operation is not. You deploy OpenReplay to AWS, GCP, Azure, DigitalOcean, Scaleway, OVHcloud, or any Kubernetes cluster, and you then own the storage, the scaling, and the upgrades. Session recordings are storage-heavy, so a busy site accumulates meaningful object storage cost, and someone has to keep the deployment healthy. For a team already running Kubernetes this is a small marginal cost; for a team that is not, it is a project.
Does OpenReplay solve my GDPR problem?
In self-hosted mode it removes the problem rather than mitigating it, because there is no third-party processor: the recordings never leave your infrastructure, so there is no transfer to assess and no vendor to review. Sanitisation happens in the tracker before transmission, so obscured content is never written anywhere. On the managed side, OpenReplay holds SOC 2 Type II, offers roughly 50 regions, and supports bring-your-own-cloud for teams who want managed operations without moving the data.
How does OpenReplay compare with LogRocket?
They are the two serious developer-facing replay tools. LogRocket is more polished, has deeper framework state integration, and adds Galileo's machine learning issue prioritisation, but is hosted in the United States on a session meter and requires an enterprise contract for EU residency. OpenReplay matches the technical category (network, console, errors, state) and adds self-hosting, open source code, and flat instance pricing. Polish versus control is the honest summary.
What are Assist and Spot?
Assist is live co-browsing: watch a user's session in real time and join a WebRTC call with them, including remote control and on-screen annotation, which lets support resolve in one conversation what would otherwise take days of email. Spot is a Chrome extension that lets anyone record a bug directly from their browser with the technical context attached and share it as a link, which dramatically raises the quality of internally filed bug reports.
Can OpenReplay replace my product analytics tool?
For a small team, sometimes. It has funnels, journeys, trends, heatmaps, and automatic issue detection over the same session data, and opening the recordings behind a drop-off is genuinely powerful. What it lacks is the analytical modelling a dedicated tool provides: behavioural cohorts, sophisticated retention analysis, experimentation, and feature flags. If those matter, pair it with Mixpanel, Amplitude, or PostHog.
Is OpenReplay a safe vendor to depend on?
It is a small company, having raised roughly $6M in total, which is a fraction of what its competitors have. The mitigation is structural rather than financial: the platform is open source on GitHub with roughly 12,000 stars, so if the company stopped tomorrow, a self-hosted deployment would keep running and the code would remain available. That is a materially better failure mode than a hosted SaaS shutting down, which is exactly what happened to Smartlook, whose end of sale came in May 2026.
Does OpenReplay work for mobile apps?
Its strength is web. The tracker, the devtools, the state integrations, and the analytics are all built around web applications, and mobile support is considerably thinner than the web story. If mobile session replay is a primary requirement, Mixpanel and Amplitude capture iOS, Android, and React Native natively and are the better fit.
Editorial verdict
OpenReplay is the answer for teams who wanted LogRocket and got told no by legal, and for teams whose traffic makes session-metered pricing absurd. Self-hosting puts the recordings inside your own cloud with no processor in the path, the managed tier charges $199 a month regardless of volume, and the technical depth (network, console, stack traces, Redux and VueX state, 40-plus performance metrics) is genuinely in the same class as the commercial leaders. Buy it if you have infrastructure competence or high traffic, and especially if data residency has previously blocked a replay purchase. Skip it if you are a marketing team wanting heatmaps, if you need mobile replay, or if nobody on your team wants to own a Kubernetes deployment, because the free tier is only free in the way that any self-hosted software is free.
Written by the SaaSTracker editorial team. Awards, when shown, are judged against the published criteria in our methodology.