Statflow

Privacy-first, cookieless web & product analytics — a self-hostable open-source alternative to Google Analytics

Own your data. No cookies. No consent banner. One command to run.

Quick start · Features · How it compares · Architecture · Privacy · Contributing

---

Statflow is an open-source analytics platform that shows how people actually use your site — not just how many visited, but where they click, how they navigate, and where they drop off. It unifies the audience measurement of Plausible with the behavioral/product analytics of PostHog and Microsoft Clarity, in a single tool you host yourself.

Because it is cookieless and never stores personal data, you do not need a cookie-consent banner — and you keep 100% of your traffic instead of losing the 20–30% that decline consent on cookie-based tools.

Status: Milestone 1 is complete and proven end-to-end — ingestion pipeline, dashboard, and tracker all run against real infrastructure with a full local test suite. See the roadmap.

✨ Features

• 📊 Audience analytics — visitors, pageviews, sessions, bounce rate, average duration, top pages, sources, devices, countries, UTM campaigns, with flexible date ranges and real-time view.
• 🖱️ Behavioral analytics — click & scroll heatmaps, rage-click and dead-click detection, engagement time, and user journeys — the why behind the numbers.
• 🔻 Funnels, retention & segments — multi-step conversion funnels, cohort retention, and saved segments that persist across reports.
• 🍪 Cookieless & private by design — visitor identity is a salted, daily-rotated hash that is never persisted; no cookies, no cross-site tracking, no consent banner required.
• 🛡️ Ad-blocker resistant — first-party proxy mode serves the tracker and ingestion endpoint from your domain, so Brave and ad-blockers do not silently drop your data.
• 🪶 Tiny tracker — the core script is under 4 KB gzipped; heavy behavioral collectors load lazily, only when enabled.
• 🏠 100% self-hosted — runs entirely on your machine with docker compose. No external CDN, no third-party API, no telemetry. Geo-IP resolves from an embedded database.
• 👥 Teams & roles — multi-user, multi-site, with owner / admin / editor / viewer roles and per-site public dashboards.

🚀 Quick start

You only need Docker (with Compose). Nothing else is installed on your machine — every tool runs in a container.

git clone https://github.com/tanguychenier/statflow.git
cd statflow
make setup

make setup is a single idempotent command: it builds the images, installs dependencies, generates local secrets, starts the data stores, runs the migrations, and brings up the API and dashboard. When it finishes:

• Dashboard → http://localhost:5173
• API → http://localhost:8001

Then add the tracker to any site you want to measure:

<script>
  window.statflowConfig = { siteKey: 'stk_your_public_site_key' };
</script>
<script src="/sf/tracker.js" defer></script>

apiBase defaults to your own origin, so the snippet works as a first-party integration out of the box. See docs/SETUP.md for production deployment and the first-party proxy.

🆚 How it compares

	Statflow	Google Analytics 4	Plausible	PostHog	Matomo
Open source	✅ AGPL-3.0	❌	✅	✅	✅
Cookieless / no consent banner	✅	❌	✅	⚠️	⚠️
Audience analytics	✅	✅	✅	⚠️	✅
Behavioral (heatmaps, journeys)	✅	❌	❌	✅	💰 paid add-on
Self-hosted, simple	✅ one command	❌	✅	⚠️ heavy	✅
Data stays 100% yours	✅	❌	✅	✅	✅

Statflow targets the gap no single tool fills today: behavioral depth + genuine data ownership + privacy by default, with self-hosting that actually takes one command. See the full competitive analysis.

🏗️ Architecture

A monorepo with a strict separation between a pure REST API and a standalone SPA, plus the tracking script:

apps/backend     Symfony 7 · hexagonal architecture · CQRS · FrankenPHP — pure REST API
apps/frontend    Vue 3 · Vite · Pinia · TanStack Query · Tailwind CSS 4 — dashboard SPA
packages/tracker Vanilla TypeScript · zero-dependency tracking script
docker/          Container images and service configuration
docs/            Architecture, ADRs, API (OpenAPI), data model, design system

Ingestion pipeline: the tracker sends a compact event → the API validates, anonymizes, and buffers it in Redis Streams → a batch worker writes it to ClickHouse → the dashboard queries aggregated metrics. PostgreSQL holds accounts, sites, and configuration.

The backend follows hexagonal architecture (ports & adapters) across five bounded contexts — Ingestion, Analytics, Identity, Sites, Reporting — with boundaries enforced in CI by Deptrac. More in docs/architecture.md and the ADRs.

Tech stack

Backend PHP 8.3 · Symfony 7 · FrankenPHP · Doctrine · Symfony Messenger Frontend Vue 3 · TypeScript · Vite · Pinia · vue-router · TanStack Query · Tailwind CSS 4 · Apache ECharts Data ClickHouse (analytical) · PostgreSQL (application) · Redis (ingestion buffer + cache) Tooling Docker · PHPUnit · Vitest · Playwright · PHPStan · Deptrac · ECS · Rector

🔒 Privacy & GDPR

• No cookies, no local storage of identifiers. Visitor and session IDs are derived server-side as HMAC-SHA256 over IP + user-agent with a daily-rotated salt that is never persisted — cross-day re-identification is cryptographically impossible after the fact.
• Cross-site isolation is built into the hash key, so the same visitor on two sites is never linkable.
• Raw IP is discarded immediately after hashing and geo-lookup; form values, query-string PII, and credentials are stripped before anything leaves the browser.
• Do Not Track and Global Privacy Control are honored by default.

Because no data is linked to an identifiable individual and no device storage is accessed, the architecture does not require ePrivacy consent. See docs/data-model/identity-and-privacy.md.

🧑‍💻 Development

Everything runs through make — nothing is installed on the host beyond Docker.

make setup     # one-command bootstrap from a fresh clone
make up        # start the stack
make ci        # run the FULL quality gate locally (lint + static analysis + tests + build)
make test      # backend + frontend + tracker test suites
make lint      # polyglot lint suite
make down      # stop the stack

CI runs locally. make ci reproduces the entire pipeline on your machine — there is no reliance on a paid cloud runner. The bundled GitHub Actions workflow is manual-only and consumes no minutes.

See CONTRIBUTING.md for the workflow, commit conventions, and how to run each gate.

🗺️ Roadmap

• Milestone 1 — Plausible parity: core audience metrics, real-time, multi-site, public dashboards, anti-adblock proxy. ✅
• Milestone 2 — GA4 parity: custom events & conversions, goals, funnels, segments, retention, roles.
• Milestone 3 — Beyond GA: full heatmaps & session journeys, data API & exports, scheduled email reports, e-commerce revenue, embeddable dashboards.

Full detail in docs/roadmap/feature-roadmap.md.

🤝 Contributing

Contributions are welcome — see CONTRIBUTING.md and the Code of Conduct. Found a security issue? Please follow SECURITY.md.

📄 License

Statflow is licensed under the GNU Affero General Public License v3.0 — see LICENSE. The AGPL keeps Statflow and every hosted derivative open source.

Copyright © 2026 Tanguy Chénier.