Bicitra Commerce
Product Architect / Lead Engineer
A self-hosted eCommerce platform: each merchant gets their own isolated Laravel/MySQL installation on shared or VPS hosting, distributed as installable release packages and kept up to date through a self-built, zero-downtime update pipeline.
- Client
- Merchants who want their own online store. My own product.
- Problem
- Store platforms charge fees and set limits. Self-hosted store software is hard to keep safely updated.
- What I built
- A store platform installed separately for each merchant, a central hub for licences and updates, and an update system that can't leave a store half-updated.
- My role
- Product Architect / Lead Engineer
- Timeline
- About a month of solo work to a working product. Launching now.
- Stack
- Laravel, PHP, Blade, Vue 3, Tailwind CSS, MySQL, Docker, Vite
Problem
Merchants who want their own store are usually stuck choosing between a SaaS platform's fees and limits, or self-hosted software that's hard to keep safely updated and hard to customize without forking core. Bicitra Commerce is built to be installed per-merchant on the cheapest available hosting, stay safely updatable long after launch, and support store-specific features without anyone touching core code.
Approach
A domain-driven Laravel 11 app: Catalog, Orders, Customers, Discounts, Payments, Shipping, Themes and Updates each isolated under app/Domain/ rather than living in fat controllers. Blade handles page structure and SEO-critical markup; Vue 3 components mount inside those views for the interactive pieces (cart drawer, gallery zoom, variant selection), deliberately with no Node.js process required in production, since assets are precompiled via Vite. A separate companion app, the "hub," acts as landlord to every individual storefront install: issuing licenses, hosting release packages, and pushing core and template updates out to stores over HTTP.
A hard problem I solved
The hard problem was shipping updates to production software running on hosts I don't control. Most installs run on cPanel shared hosting, where symlink() can be disabled outright and file ownership can end up inconsistent after a manual composer install, with no ops team to step in if an update goes wrong. The update pipeline now preflights every file write before the site goes down, so a permission problem is caught and reported with zero files touched instead of discovered mid-update; verifies each release package against a checksum before applying it; skips replacing vendor/ entirely when dependencies haven't changed; and rolls back completely on failure, including removing new top-level files a failed update had already added. That preflight check exists because of a real case: a host where vendor/ had been left owned by the SSH user after a manual composer install. I first found it the hard way, mid-update, with the site down.
Status
Currently launching. Built solo from the initial commit through a working self-update pipeline, a two-theme storefront, and a sandboxed extension system (extensions boot inside their own try/catch and auto-disable after three consecutive failures, so a bad one can't take a storefront down) in about a month of focused engineering. Not yet at scale: the safety mechanisms were built ahead of the first paying merchant, not retrofitted after something broke.
Technology
- Laravel
- PHP
- Blade
- Vue 3
- Tailwind CSS
- MySQL
- Docker
- Vite
Writeup
Hiring a developer? Get the checklist first
20 checks founders can run before they sign: questions to ask on the first call, who should own the code and accounts, and what senior remote developers cost in 2026. A free two-page PDF, plus three short emails from real builds. Unsubscribe any time.
Need something similar built?
I build and run products like this for founders and businesses worldwide, from architecture to deployment.
Or tell me about your project →Related: DevOps for Startups