Guroos — AI-Integrated Author Marketplace

Overview

Guroos connects readers, listeners and viewers with the people behind the ideas they care about. A reader who finishes a book, podcast episode or lecture can find the author or expert, see transparent per-session pricing and availability, and book a private one-on-one video call.

The product is a two-sided marketplace: authors and experts manage profiles, availability, sessions and earnings; users discover experts, book and pay, and continue the conversation through private messages.

The business need

The client needed more than a booking page. A marketplace has to handle two very different kinds of users, move money between them fairly, and stay manageable for a non-technical team.

• Content operations had to remain simple, so editorial content and profiles needed a familiar CMS.
• The user experience had to feel like a modern product — fast search, filtering and booking — not a traditional CMS theme.
• Payments had to split automatically between the expert and the platform, with expert payouts handled properly.
• Profile creation is a cold-start problem: experts often arrive with little written about themselves.

My role

I was responsible for the engineering of the platform end to end: the headless architecture, the Next.js frontend, the WordPress data model and REST endpoints, the AI-assisted content workflow, the payment integration, real-time messaging, and the role-specific dashboards.

Technical architecture

Guroos uses a headless architecture. WordPress acts as the content and data backend, exposed through the WordPress REST API, while a Next.js and React application written in TypeScript delivers the entire user-facing experience.

• Frontend: Next.js, React, TypeScript and Tailwind CSS, with separate experiences for users, experts and administrators.
• Content and data: headless WordPress with custom REST API endpoints, backed by MySQL.
• Real-time layer: a Node.js service using Socket.IO for private messaging.
• Payments: Stripe Connect for marketplace payments and expert payouts.
• AI services: the OpenAI API for author-profile and content generation.
• Infrastructure: deployed on DigitalOcean.

Key implementation decisions

Headless WordPress instead of a custom CMS. WordPress gives the client a mature, familiar editing environment and a stable REST API, while Next.js provides the product-grade interface. This kept the operational burden low without constraining the frontend.

Generative AI as an assistant, not an author. OpenAI-assisted generation is used where it removes friction — helping produce author-profile and supporting content — rather than as a headline feature.

Commission handled by the payment platform. Using Stripe Connect means the 15% platform commission is applied automatically as part of each payment flow, and expert earnings and withdrawals follow Stripe's connected-account model rather than manual reconciliation.

A dedicated real-time service. Messaging runs on Socket.IO in Node.js, separate from the request/response WordPress API, so live conversations do not depend on CMS request cycles.

Key features

• Author and expert discovery with search and filters by topic, expertise, rating and availability
• Public expert profiles with transparent per-session pricing
• Booking flow with an availability calendar
• Stripe Connect payments with an automatic 15% platform commission
• Expert earnings tracking and withdrawal workflows
• Private real-time messaging between users and experts
• OpenAI-assisted author-profile and content generation
• Role-specific dashboards for users, experts and administrators

Delivery status

The platform is production-ready. The client has not yet officially launched it, so this case study makes no claims about traction, user numbers or commercial results.