LinkBay-CMS: the story of an e-commerce infrastructure built for agencies

LinkBay-CMS is a multi-tenant e-commerce platform for people stuck between the code and the client: agencies, software houses, brands with several stores.

It's not for the shop owner who wants to open a store. It's for the agency that runs all its clients' stores from one dashboard, under its own brand.

The tagline: "Built for the agency, not the merchant. The merchant is your client. The platform is yours."

Every time an agency takes on a new e-commerce client, the same thing happens: either they rebuild stuff that already exists, or they glue together tools that don't talk to each other.

Agencies are good at understanding clients. They shouldn't have to be sysadmins for each one.

E-commerce platforms are built for the merchant, not for whoever supports them. Managing 10, 20, 50 stores? With digital duct tape. And tools like ManageWP, MainWP and WP Umbrella do backups and updates, they don't help you sell infrastructure as a service.

Hence the idea: a platform for the agency, not the merchant.

Easy to say, hell to build: one dashboard for tenants, branding and stores, with enough automation that you don't start from zero with every client. Reusable layouts, blocks, themes and features. Brand, prices and rules are the agency's call.

The business model: the platform takes a public, transparent cut of what clients pay the agency. No "custom enterprise pricing", which translates to "I'll charge whatever I can get away with".

If the agency grows, LinkBay grows. If it stands still, I don't earn. Seems fair.

I design, write and maintain all of it. Every line of code, every architecture choice. Not to show off: the problem is real and existing tools didn't solve it.

The stack:

• multi-tenant backend with Laravel and PostgreSQL, Redis for caching;
• JWT, CSRF, sessions and RBAC;
• Stripe for payments and subscriptions;
• site in Next.js, React, TypeScript and Tailwind;
• Docker, Traefik and Node-RED for automations.

In about 21 months I went from the foundations (login, tenants, billing) to an actual product. Between spring and summer 2026 the big pieces landed:

• Layout Manager: reusable layouts, snapshots and rollback, shared blocks, live preview;
• Plugin System: central registry, hot-reload, sandboxed plugins;
• Marketplace and licensing: each agency pays only for what it uses;
• Theme Engine: themes, live preview, CSS generated from variables, so every client gets its own look without duplicating code.

Linear? Not a chance. There were months when it looked stuck. It wasn't: I was fixing things under the hood.

In mid-2026 I had a look in the mirror: building a do-everything mega SaaS alone isn't realistic. Let's be honest.

Some agencies had confirmed the problem is real: managing lots of clients from one place. They hadn't asked for a giant system.

So I cut. Less architect ego, more of what actually matters: reliable feature-gating, reusable components, a consistent backend. No complicated upsells, no public marketplace to run.

Better a product I can maintain than a beautiful, fragile architecture.

On July 21, 2025 I filed "LinkBay-CMS" with the UIBM (Italian Patent and Trademark Office). On June 3, 2026 the certificate arrived: registration number 302025000116815.

Three classes:

• 42: SaaS, cloud, software development and maintenance;
• 9: computer programs, management and multimedia software;
• 38: data transmission and online chat.

It's valid for ten years from filing. It's not just paperwork: the name I use with agencies is now mine on paper too.

LinkBay-CMS is live at linkbay-cms.com: a "control deck" to run all your clients' stores, under the agency's brand, with configurable subscriptions and a public platform fee for each plan.

It's not done. Actually, the selling part has just started. But compared to a year ago it has a registered name, a solid base and proof that the problem actually exists.

If you work at an agency: how many stores are you holding together with duct tape?