When you launch a web application or micro-SaaS, understanding how users actually interact with your product is non-negotiable. You need to know which onboarding steps cause users to drop off, which buttons get clicked, and how retention cohorts perform over thirty days.
For years, founders defaulted to Google Analytics 4 or commercial event trackers like Mixpanel and Amplitude. But GA4 is bloated, confusing, and privacy-hostile, while Mixpanel quickly ramps up into steep per-event monthly bills once your application gains meaningful user traction.
That tension is what propelled PostHog (github.com/PostHog/posthog) into one of the most celebrated open-source developer platforms in tech.
PostHog promises an entire product analytics suite in a single platform: event auto-capture, session recordings, feature flags, A/B experimentation, user heatmaps, and SQL survey funnels.
I wanted to evaluate PostHog strictly through the lens of an independent solo bootstrapper: what does it actually take to self-host this repository, what are the hidden infrastructure costs, how does the licensing work for commercial ventures, and should a solo operator run it locally or use their hosted cloud tier?
Here is the realistic technical assessment.
Repository Snapshot (Data verified September 17, 2026)
- GitHub Repository: https://github.com/PostHog/posthog
- Stars: 39,837 | Forks: 3,393 | Open Issues: 5,530
- License: MIT (Core) + PostHog Enterprise License (Feature Flags & Multi-tenancy)
- Primary Language: Python & TypeScript (ClickHouse, PostgreSQL, Kafka, Redis)
- Contributors: ~450+ | Last Commit: September 17, 2026
The Reality of Self-Hosting: The ClickHouse & Kafka Ceiling
If you look at PostHog’s GitHub repository, you will see instructions for running a Docker container or Kubernetes Helm chart.
For a software engineer, spinning up a local Docker compose environment looks simple. But running PostHog reliably in production for a live commercial business is an entirely different beast.
Under the hood, PostHog is not a lightweight SQLite or Postgres app. To handle high-volume event ingestion without locking up queries, PostHog relies on an industrial-strength data pipeline:
- ClickHouse: An ultra-fast columnar analytical database engine.
- Apache Kafka: A distributed streaming platform to buffer incoming event telemetry.
- Redis: For caching and job queuing.
- PostgreSQL: For core user records and application settings.
Incoming User Events -> Load Balancer -> Kafka Buffer -> ClickHouse Engine -> PostHog UI
(Requires minimum 4 vCPUs, 16GB RAM, and persistent high-IOPS NVMe storage)If you attempt to self-host PostHog on a cheap $10/month DigitalOcean droplet or small VPS, the containers will run out of memory (OOM) within hours.
To run PostHog smoothly in production without query timeouts, you need a server with at least 8 vCPUs and 16GB of RAM, plus continuous disk backups for your ClickHouse analytical tables. On AWS or Hetzner, that infrastructure baseline costs $60 to $120 per month before accounting for your personal maintenance labor.
When ClickHouse corrupted table partitions or Kafka broker rebalances occur, you are the on-call DevOps engineer debugging analytical databases at 11:00 PM instead of talking to your customers.
Licensing: What the MIT vs Enterprise Mix Permits
PostHog uses a hybrid open-source model.
The core data ingestion and event analytics engine is licensed under MIT, granting complete freedom to inspect, modify, and run the code for private commercial products without paying royalties.
However, advanced enterprise capabilities—such as automated user identity stitching, team role-based access control (RBAC), and specific multitenant feature gating—are gated behind proprietary license terms. For a solo founder or small startup building their own product, the open-core MIT features provide more than enough analytical fidelity.
The Solo Founder Verdict: Self-Host or Cloud?
Here is my direct recommendation as a solo operator:
Do not self-host PostHog if you are a solo founder with fewer than 100,000 monthly active users.
PostHog offers a remarkably generous PostHog Cloud Free Tier:
- 1,000,000 events per month completely free.
- 5,000 session recordings per month completely free.
- Unlimited feature flags and A/B test splits.
Managing a distributed Kafka and ClickHouse cluster yourself to save $0 on an analytics bill is peak founder distraction. By using their hosted cloud tier, you get the world-class PostHog UI, flawless session playback, and instant setup via a lightweight JavaScript snippet, leaving your infrastructure time focused entirely on your core product.
If you strictly require 100% self-hosted privacy compliance with zero third-party data transmission, look at lighter alternatives like Umami Analytics (github.com/umami-software/umami) or Plausible, which run effortlessly on a $5 VPS.
For more infrastructure breakdowns, read our hands-on review of Coolify: the self-hosted PaaS for solo founders and our architectural evaluation of PocketBase as a solo founder backend.