OpenPanel
Open-source web and product analytics in one tool
OpenPanel is an open-source analytics platform that combines web analytics with product analytics, positioned as a mix of Mixpanel and Plausible. It tracks events, funnels, and user retention.
Key features
- Web and product analytics
- Funnels and retention
- Real-time event tracking
- Privacy-friendly design
Pros & cons
Strengths
- Mixpanel-style features open-source
- Modern dashboard
Trade-offs
- Newer project
- Multi-service deployment
OpenPanel replaces
Last reviewed Sep 13, 2026 · 771 words
OpenPanel is a Mixpanel replacement that also does web analytics, not a Plausible replacement that also does events. If what you want is a pageview counter with referrers and a bounce rate, deploy Plausible or Umami and be done in 20 minutes. OpenPanel's extra moving parts buy you funnels, retention cohorts and per-user event timelines, and that price is only worth paying if you will actually sit down and read a funnel.
Six containers where Umami needs two
The self-hosting layout is an API service, a dashboard, a worker, PostgreSQL for metadata, ClickHouse for events and Redis for queues. The repository ships a self-hosting directory with a setup script that writes the environment file and the compose file for you, and that is the path to take; hand-assembling it from the individual images is possible but is how people end up with mismatched versions. Give it a hostname and a certificate through your usual reverse proxy and the script does the rest.
The listing's 512 MB is the floor, and it is a floor for a hobby site. ClickHouse is the component that decides your memory budget, and it is content at 2 GB and unhappy below 1 GB once the events table has a few million rows. Plan for a 2 to 4 GB box if the site you are tracking has real traffic.
Events are the unit; pageviews are a special case
The web snippet auto-tracks screen views, referrers, devices and countries the way any analytics script does. The difference is what else you send:
op.track('signup_completed', { plan: 'pro', source: 'pricing_page' });
op.identify({ profileId: user.id, email: user.email });
Once events carry properties and a profile, you get the Mixpanel toolbox: a funnel from pricing_viewed to signup_completed with drop-off per step, retention by weekly cohort, and a per-user timeline showing every event one account fired. SDKs exist for plain JavaScript, React, Next.js, React Native and Node, so the same event stream can come from a marketing site and the app behind it. That combined view, one tool for the public site and the logged-in product, is OpenPanel's actual pitch and it is a real gap in the self-hosted options.
Where it sits against PostHog
PostHog is the bigger project and the safer bet if you need session replay, feature flags or A/B experiments, none of which OpenPanel has. The trade is weight: PostHog's self-hosted footprint is larger still, and the company has openly steered self-hosting toward hobby use, with paid cloud as the supported path. OpenPanel is the middle option: more than web analytics, less than a full product-engineering platform, and still deployable by one person on one machine. If you already know you need replay, skip the middle.
Privacy holds until you call identify
Out of the box OpenPanel is cookieless and stores nothing that identifies a person, which keeps it in the same consent-free position as Plausible for a public site. The moment you call identify with an email, you are storing personal data on a server you operate, with all the retention and access obligations that implies. That is not a flaw; it is the difference between web and product analytics. Just decide which mode a given property is in before the data arrives, because the decision to leave Google Analytics is usually made on privacy grounds, and a product-analytics setup can quietly undo it.
Two years old is young for a data store
The project dates from 2024. The dashboard is polished and development is quick, and that speed cuts both ways: expect schema migrations between versions, read release notes before pulling latest, and keep ClickHouse and Postgres on volumes you back up. There is a hosted cloud from the same team, which is a reasonable sign that the company intends to survive, and an export path through the API for the day you want your events somewhere else.
What I'd do
For a personal site, a blog or a documentation host, use Plausible or Umami and give it 20 minutes; the Plausible versus Umami comparison is the fork to read. For a product you are building, where you would otherwise be paying for Mixpanel, run OpenPanel on a 4 GB box, wire the web snippet into the marketing site and the SDK into the app, and define 5 to 10 named events on day one rather than tracking everything. Revisit PostHog only when someone asks for session replay, and not before.
Compare OpenPanel
22 head-to-head comparisons.
- OpenPanel vs PostHog
- OpenPanel vs Umami
- OpenPanel vs Plausible
- OpenPanel vs Matomo
- OpenPanel vs Rybbit
- OpenPanel vs Fathom Lite
- OpenPanel vs GoatCounter
- OpenPanel vs Countly Community Edition
- OpenPanel vs Ackee
- OpenPanel vs Shynet
- OpenPanel vs Tianji
- OpenPanel vs Open Web Analytics
- OpenPanel vs Counterscale
- OpenPanel vs Vince Analytics
- OpenPanel vs Aptabase
- OpenPanel vs Litlyx
- OpenPanel vs Swetrix
- OpenPanel vs Pirsch
- OpenPanel vs Counter.dev Selfhost
- OpenPanel vs Offen Fair Web Analytics
- OpenPanel vs Rakam Analytics
- OpenPanel vs Medama
Similar web analytics apps
Superset
Web AnalyticsModern data exploration and visualization platform
Replaces Tableau, Power BI
Metabase
Web AnalyticsOpen-source business intelligence and dashboards
Replaces Tableau, Looker
PostHog
Web AnalyticsOpen-source product analytics platform
Replaces Google Analytics, Mixpanel
Umami
Web AnalyticsSimple, fast, privacy-focused web analytics
Replaces Google Analytics
Postiz
Web AnalyticsSchedule posts, track the performance of your content, and manage all
Replaces Buffer, Hootsuite
Netron
Web AnalyticsVisualizer for neural network and machine learning models