Zenovay: Real time analytics platform built on Supabase + Cloudflare Workers

[jasko-ctrl]

Zenovay — Real-time analytics platform built on Supabase + Cloudflare Workers

We built Zenovay, a privacy-first web analytics platform with a real-time 3D globe that shows website visitors arriving live.

Supabase is the backbone of the entire product.

What we built

A web analytics tool that answers "which marketing channels bring paying customers?" instead of just "how much traffic did we get?" Real-time tracking, AI-powered visitor scoring, revenue attribution through Stripe, and a 3D globe visualization powered by Mapbox GL JS.

EU-hosted, cookieless by default, GDPR compliant without consent banners.

How we use Supabase

Database

• 143 tables in public schema
• 119 foreign keys
• RLS enabled on every single table
• Extensions: pg_cron, pg_net, uuid-ossp, pgcrypto

Auth

• Supabase Auth with JWT + iron-session cookies
• Email/password, Google OAuth, GitHub OAuth
• SSO (SAML/OAuth) for enterprise customers
• MFA (TOTP) and WebAuthn passkeys

Realtime

• Live subscriptions on visitors and analytics_daily tables
• Powers the real-time visitor count and the 3D globe markers appearing in real-time
• When a visitor lands on a tracked site, Cloudflare Workers process the event at the edge in under 50ms, write to Supabase, and Realtime pushes it to the dashboard instantly

pg_cron + pg_net (scheduled jobs)

We run a bunch of background jobs directly in Postgres:

Schedule	Job
Hourly	Subscription expiration warnings, auto-downgrade
Every 2 hours	Traffic spike alerts
Daily	Retention notifications, ticket auto-close, dunning (card expiry, reminders, enforcement)
Monday 9AM	Weekly summary emails

All triggered via pg_cron calling our API endpoints through pg_net HTTP calls.

What we store

Core analytics (visitors, events, pageviews), pre-aggregated daily stats, AI-generated insights, conversation history, knowledge base embeddings, heatmap data, session recordings, conversion funnels, support tickets, community forum posts, billing/subscription state, and more.

Architecture overview

Browser --> Cloudflare Workers (edge, sub-50ms) --> Supabase PostgreSQL (cold storage)
                                                --> Cloudflare KV (hot cache)
                                                --> Cloudflare Analytics Engine (high-cardinality events)
                                                          |
                                               Supabase Realtime --> Dashboard (Next.js)

The edge-first approach means our tracking script adds zero measurable latency to the sites it monitors. Events are processed at the nearest Cloudflare data center (300+ locations), then persisted to Supabase.

Tech stack

Layer	Tech
Edge API	Cloudflare Workers + Hono.js
Database	Supabase PostgreSQL
Auth	Supabase Auth
Realtime	Supabase Realtime
Caching	Cloudflare KV
Object Storage	Cloudflare R2
Live Visitors	Cloudflare Durable Objects
Event Queries	Cloudflare Analytics Engine
Dashboard	Next.js 14 (App Router)
3D Globe	Mapbox GL JS
Payments	Stripe
AI	Cloudflare Workers AI + OpenAI via AI Gateway

Numbers

• 16 repositories, all deployed on Cloudflare Pages/Workers
• 7 static sites auto-deploying from GitHub
• Sub-50ms event ingestion globally
• 99% tracking success rate (including Safari Private Mode at 95%)

Links

• Live: https://zenovay.com
• Docs: https://docs.zenovay.com
• Architecture deep-dive: https://dev.to/zenovay/how-we-built-a-real-time-analytics-platform-on-cloudflare-workers-architecture-deep-dive-5h3h

Happy to answer any questions about how we're using Supabase or the architecture in general.