Overview
Self-hosted product analytics with session replay, feature flags, experiments, and error tracking.
PostHog is a self-hosted product analytics platform for teams that want analytics, session replay, feature flags, experiments, surveys, and error tracking in one system.
Why people choose PostHog
PostHog is broader than a simple website-statistics dashboard. Its official product pages group product analytics, web analytics, session replay, feature flags, experiments, surveys, and error tracking into one developer-focused toolkit. That makes it relevant when a team wants to connect what users do with how the product behaves.
The self-hosted option is also a data-control choice. Events and product data run on infrastructure you operate rather than being sent to a third-party analytics vendor by default.
What you get in the stack
Core capabilities include event-based product analytics, web analytics, session recordings, feature flags, A/B testing, surveys, error tracking, heatmaps, and data pipelines. The exact mix and operating cost depend on the deployment and workload.
PostHog also provides SDKs and APIs for instrumenting applications, while its source repository and official documentation cover the hobby deployment path and operational tooling.
What to know before you deploy
Self-hosted PostHog is not a small single-container homelab service. The official self-hosting guide describes a Linux Ubuntu VM equivalent to 4 vCPU, 16 GB RAM, and more than 30 GB of storage for the hobby deployment, plus a domain and DNS A record.
PostHog states that self-hosted deployments are officially unsupported, do not receive the same guarantees as PostHog Cloud, and should stay on the latest Docker image because changes ship continuously. You own upgrades, backups, scaling, and security response.
Choose the self-hosted path when data control and infrastructure ownership justify that work. For a small site that only needs lightweight traffic reporting, a simpler tool such as Plausible or Umami may be a better operational fit.
Best fit
PostHog fits product teams, developers, and operators who need analytics tied to releases, experiments, feature rollout, and debugging. It is a poor fit when the goal is only a tiny low-maintenance pageview counter or when you need vendor-backed support guarantees.
Interface previews
Screenshots of the PostHog interface.
Feature support
| Feature | Support |
|---|---|
| Product analyticsOfficial product documentation describes event-based product analytics for understanding user behavior. | Supported |
| Web analyticsPostHog lists web analytics as a separate product capability. | Supported |
| Session replaySession recordings are a core PostHog product area; privacy and retention settings remain operator responsibilities. | Supported |
| Feature flagsFeature flags support controlled rollout and product experimentation. | Supported |
| A/B testing and experimentsPostHog provides experiments, including no-code web experiments in its documentation. | Supported |
| SurveysSurveys are listed among the official product capabilities. | Supported |
| Error trackingError tracking is part of the developer-tool stack described by PostHog. | Supported |
| HeatmapsHeatmaps are listed as a PostHog product capability. | Supported |
| Self-hosted Docker deploymentThe official hobby deployment is available, but self-hosting is officially unsupported and requires substantial infrastructure. | Partial support |
| Kubernetes deploymentThe self-hosting guide says new deployments of the paid open-source product using Kubernetes are no longer supported. | Partial support |
Community signals
- Stars
- 40.3k
- Forks
- 3.5k
- Open issues
- 5.7k
- Last commit
- 22h ago
- Latest release
- desktop-v0.61.729
- Repo created
- Jan 2020
Includes open pull requests
1d ago
Refreshed nightly from the GitHub API.
Alternatives
- Plausible Analytics
Simpler privacy-focused analytics comparison for sites that do not need the full product stack.
- Umami
Lighter self-hosted web analytics comparison.
- Matomo
Broader analytics comparison with a different deployment and reporting model.
Questions
What is PostHog used for?
PostHog is used for product and web analytics, session replay, feature flags, experiments, surveys, heatmaps, and error tracking.
Can PostHog be self-hosted?
Yes. PostHog publishes a free Docker Compose hobby deployment, but its official documentation says self-hosted instances are unsupported and require you to manage the infrastructure.
What are the server requirements for self-hosted PostHog?
The official guide describes an Ubuntu Linux VM equivalent to 4 vCPU, 16 GB RAM, and more than 30 GB storage, plus a domain with an A record. Workload growth can require more.
Is self-hosted PostHog free and open source?
PostHog is open source and provides a free self-hosting deployment under an MIT-licensed path, but infrastructure, operations, and scaling are still your responsibility.
Is self-hosted PostHog supported like PostHog Cloud?
No. PostHog explicitly says self-hosted deployments are officially unsupported and do not come with guarantees about how the product works on your infrastructure.
Should I use PostHog or a simpler analytics tool?
Use PostHog when you need product analytics connected to flags, experiments, replay, and errors. For basic privacy-friendly website statistics with less infrastructure, compare Plausible or Umami.
How does PostHog self-hosting handle updates?
The official guide recommends staying on the latest Docker image because self-hosted changes ship continuously. Back up data before upgrades and monitor the deployment yourself.
support // the lab
Found this write-up useful?
If it saved you time or a rebuild, you can support more practical homelab guides.