PostHog is the open-source product-analytics platform most teams compare with Mixpanel and Amplitude. It does the same jobs that Mixpanel prices by event volume (events, funnels, retention, cohorts) and adds session replay, feature flags and experiments. Its core is MIT-licensed, and it can be self-hosted, with limits you should know before you start.
Short answer
PostHog covers what most teams buy Mixpanel or Amplitude for. Code outside its ee/ directory is MIT-licensed, and PostHog publishes a free Docker Compose deployment for self-hosting. PostHog itself says its Cloud is the better experience for most users: self-hosted instances get no support from PostHog, and all paid-plan features are Cloud-only. It suits engineering-led teams who want their event data; less technical teams usually do better on a hosted tool.
I run product-analytics and event-pipeline work at MithTech, a team of Frappe specialists in Bengaluru. Founders ask how to understand user behaviour without sending every click to a US vendor they cannot audit, and PostHog is where we usually start. PostHog is a strong product. Self-hosting it is a real infrastructure commitment, not a weekend toggle.
What is PostHog, and how does it differ from Mixpanel and Amplitude?
PostHog captures user events and turns them into funnels, retention curves and cohorts, the same job Mixpanel and Amplitude do. Unlike those two, it publishes its code so you can read it, self-host it and control the data end to end.
Mixpanel and Amplitude are polished, cloud-only, closed-source products. You send them your events and they store and analyse them on their infrastructure. That works until you need data inside your own boundary, need to audit how it is processed, or want off a plan that grows with your volume.
PostHog changes that. With the MIT core you can run it on a server you own and keep raw event data in your own environment. For Indian businesses that care about data residency, or about not renting back their own product data, that matters. How we set up analytics for clients is on the analytics and BI page.
Is PostHog really open source and self-hostable?
Yes, with conditions. The repository's licence file puts everything outside the ee/ directory under MIT, which covers event capture, ingestion, the API and the standard analytics views. Code in ee/ is under a separate licence, as in most open-core projects. If you want no proprietary code at all, PostHog maintains the posthog-foss mirror.
Self-hosting is real. PostHog documents a free Docker Compose deployment, installed and upgraded with its hobby scripts, on a Linux server you control; point your app's tracking at it and events never leave your infrastructure.
The conditions, from PostHog's own self-hosting page (October 2026):
- It suggests a machine roughly equivalent to 4 vCPU, 16 GB RAM and 30 GB+ of storage.
- It does not offer support, paid or free, for self-hosted instances.
- All paid-plan features are Cloud-only.
- It says PostHog Cloud is "far and away the best experience" for most users, and its Cloud free tier includes 1 million events a month.
So "self-hostable" is true, but PostHog treats it as a starting point. For a high-traffic app, scaling the underlying data stores is hard work you would own alone.
What does PostHog do that Mixpanel and Amplitude also do?
It captures events, builds conversion funnels, measures retention, groups users into cohorts and lets you query behaviour with a flexible interface. Those are the capabilities most teams buy Mixpanel or Amplitude for.
PostHog also bundles tools that are usually separate purchases:
| Product analytics (events, funnels, retention) | Mixpanel, Amplitude, Heap | Yes |
| Session replay (web and mobile) | Hotjar, FullStory | Yes |
| Feature flags and A/B tests | LaunchDarkly | Yes |
| Surveys | Survey tools | Yes |
| Error tracking | Sentry-style tools | Yes |
| Web analytics | Google Analytics | Yes |
Session replay is the standout for debugging real user behaviour. PostHog records sessions with console logs, network activity and a DOM view, and you can filter recordings by user, event or feature flag. Numbers from funnels and replays of the same sessions in one tool would otherwise take two or three vendors.
Where does PostHog fall short?
Usability and operational scale, not features. PostHog is built by engineers for engineers, and the interface shows it. Product managers, marketers and analysts often find it harder to learn than Mixpanel or Amplitude, whose interfaces are tuned for non-technical daily use. If your main analytics users are not developers, adoption can stall outside engineering.
The second gap is the self-hosted edition, covered above: no support, paid features on Cloud only, and scaling work that is yours. And because PostHog does so much, there is more to learn than in a single-purpose tool. If what you need is website traffic rather than product events, a lighter tool fits better; see the open-source Google Analytics alternatives or how to self-host Plausible.
Who should choose PostHog over Mixpanel or Amplitude?
Teams where engineering owns analytics and wants to control the data. Teams that want raw events in their own environment, prefer an open licence, and can run infrastructure, or are happy on PostHog Cloud.
Choose PostHog if
You are engineering-led, want to own your event data, would rather combine analytics, replay and flags in one tool, and can either self-host or use the Cloud free tier.
Consider Mixpanel or Amplitude if
Analytics is owned by non-technical product or growth people who need a gentle interface, you don't want to run infrastructure, and closed-source cloud SaaS fits your data policy.
For most early-stage Indian businesses we work with, the practical path is PostHog Cloud's free tier first, and self-hosting later if data residency or cost justifies the operational work.
How do you get started with PostHog?
An evening on the hosted tier, or a planned infrastructure task if you self-host.
Decide hosted or self-hosted
Start on PostHog Cloud if you want speed and no operations, or self-host if keeping data on your own infrastructure is a hard requirement. Be honest about which one your team can support without vendor help.
Deploy or sign up
For self-hosting, provision a Linux server of the size PostHog suggests and run its Docker Compose deployment. For Cloud, create an account.
Instrument your events
Add PostHog's snippet or SDK to your web or mobile app and define the events that matter: sign-ups, key actions, conversions. Clean event design up front is what makes funnels and retention useful later.
Build funnels and turn on replay
Create your first funnels and retention views, and enable session replay to watch how real users move through the flows your funnels measure.
Review fit and scale
Once data flows, check how the interface suits your team and watch event volume. If a self-hosted instance grows past what you can run comfortably, plan a move to Cloud.
Frequently asked questions
+Is PostHog free?
The MIT-licensed core is free to self-host, and PostHog Cloud has a free tier (1 million events a month, per PostHog's page in October 2026). Paid plans cover higher usage and extra features; check PostHog's pricing page for current figures.
+Can PostHog fully replace Mixpanel and Amplitude?
For most teams, yes: it covers events, funnels, retention and cohorts, and adds session replay and feature flags. The main trade-off is the interface, which non-technical users find steeper than Mixpanel's or Amplitude's.
+Is self-hosting PostHog difficult?
Starting is straightforward with Docker Compose. Running it in production is a real commitment: databases, upgrades, backups and monitoring, on a machine of about 4 vCPU and 16 GB RAM, with no support from PostHog and paid features only on Cloud.
+Does PostHog include session replay?
Yes, for web and mobile, with console logs, network activity and a DOM view. You can filter recordings by user, event or feature flag.
+How is PostHog different from Google Analytics?
Google Analytics focuses on aggregate website traffic and marketing metrics; PostHog focuses on what individual users do inside a product: events, funnels, retention and replay. PostHog also has web analytics, so some teams use it for both. For website stats alone, see the Google Analytics alternatives.
+What licence is PostHog released under?
The repository is MIT-licensed except for the ee/ directory, which has its own licence. PostHog also publishes posthog-foss, a mirror with the ee/ code removed, for teams that want only free software.
If you want help choosing between PostHog Cloud and a self-hosted instance, or wiring it into the rest of an open-source stack, we can help.
Comments & ratings
No comments yet. Start a new discussion.
Manoj · Published on 20 July 2026
Worked with us? Leave a Google review