Weplex engineering - Weplex3 means routes are data

Mountain View and Ribeirão Preto, Brazil — August 16, 2025. For some time, our platform carried too many handwritten route rules inside the codebase. This made sense at first, but as WePlex grew it became clear that routes themselves are part of the content domain. If WePlex is about generating websites from structured specifications, then route definitions also belong in the database alongside the templates.

The complication was that every new client type or informational page meant duplicating boilerplate in routes-fat.js, with repetitive fetch blocks for SSR and CSR. The question became: how do we unify these definitions so that routes can evolve as data, not hard-coded logic?

The answer arrived in WePlex3. We introduced JSON page specs with an embedded route section. Each spec now carries its own path, component mapping, and query list. Our codegen scripts (gen-page.js, gen-all.js) generate the component shells, and a new gen-routes.js aggregates the route specs into routes_generated.js. This file exports both the normalized route metadata and the necessary imports, which we merge into the main routes array. The order of routes is preserved so catch-all logic still functions.

Along the way, we simplified the handling of informational pages like privacy policy and terms of use, reducing them to straightforward slice definitions instead of gallery scaffolding. The build process now includes a single npm run gen:full that regenerates all templates and routes in one pass.

WePlex3 therefore consolidates routes into the same generation pipeline as components, ensuring that what used to be handwritten glue is now fully spec-driven.

On behalf of CEO Marcio S Galli, this is the launch note for WePlex3.