Overview

WordPress and Shopify sites connect to the Jorge Castro SEO platform through plugins. Every other stack, such as Laravel, Django, Rails, Next.js or a home-grown CMS, uses the Custom Integration. The customer implements a small REST contract on their own backend, and when a user clicks Publish, Update or Delete in the platform, the platform pushes signed HTTPS requests to those endpoints. This project is the public developer documentation for that contract, built with Mintlify and backed by two OpenAPI specifications.

My role

I built this integration end to end. I wrote the custom-platform adapter in the platform backend, a zero-dependency reference receiver for end-to-end testing, and this documentation site with both OpenAPI specs, the guides and the sample implementations.

What it does

• A quickstart and guides for blog posts, products, categories, authors, SEO metadata and page sync
• Authentication docs with request-signature verification samples
• Testing, error, troubleshooting and go-live checklist pages
• Reference implementations in Node.js, Laravel, plain PHP and Python
• An API reference generated from the contract spec and the platform's own connection, status and log endpoints

How it's built

The docs are MDX pages in a Mintlify project with a custom theme, and the API reference is generated from two OpenAPI files: one for the endpoints a customer implements, one for the platform endpoints they call. On the platform side, the adapter is a Node.js module that sits beside the WordPress and Shopify publishers and returns the same normalised response shape, so the existing publishing dispatchers use it without changes.

Highlights

• Only three endpoints are required: a handshake, create post and update post. During the handshake the site declares optional capabilities, and the platform enables matching features in its UI.
• Requests carry an HMAC-SHA256 signature over a timestamp and the raw body, with verification code in four languages.
• Update endpoints use partial-update semantics, so a content-only or SEO-only push never needs to read the record back first.
• The reference receiver also serves public preview pages, so the platform's crawler can test page sync against it like a real site.