🏆 MachineTalents Hackathon — Winning Project (July 2025)

This repository contains our winning project from the MachineTalents Hackathon, held in July 2025.

We achieved 1st place in the MCP category — see the announcement on LinkedIn.

---

👥 Team

• Andrei Lahonou
• Raj P
• Tom Nash

---

Claude example with mcp server:

Cline example with mcp server

Objective

Develop a Model Context Protocol (MCP) server using SUBJCT's provided API endpoints to demonstrate capabilities and solve real-world content optimisation challenges via an agentic delivery method.

See: https://modelcontextprotocol.io/introduction

Scope

• Use Subjct's API endpoints to build a functioning MCP server within the 4-hour timeframe.
• Focus on transforming content from a CMS (e.g., Shopify or WordPress) for optimisation:
  • Add relevant links, tags, and schema.
  • Integrate knowledge graph data.
  • Ensure compatibility with existing SUBJCT functionalities.

Requirements

• Utilise provided Swagger documentation and test account access.
• Deliver a proof of concept showcasing seamless API integration.
• Highlight how the solution can optimise articles for increased LLM mentions.
• Show the MCP server plugged in to Claude, ChatGPT or Google ADK to demonstrate functionality

Resources Provided

• Test account credentials. Sign up using this link.
• SUBJCT API documentation (Swagger): https://api.subjct.ai/docs/swagger/
• After signing up, access the keys in the config area.
• MCP Server Quickstart
• Using LLMs to generate MCP server
• MCP + Google ADK

Deliverables

• A working MCP server prototype.
• Documentation and a brief presentation demonstrating functionality.
• Suggestions for scaling the solution or next steps.

Evaluation Criteria

• Functionality and completeness of the MCP server.
• Effective utilisation of SUBJCT API.
• Potential for real-world application and scalability.
• Creativity and problem-solving approach.
• Going the extra mile to illustrate the power of MCP servers when bundled together (e.g. a demo illustrating your SUBJCT MCP server combined with Shopify/Wordpress CMS MCP server within Claude/ChatGPT)

Development Notes

Example Article for Testing

Use the following article for testing: https://emmajaneknight.com/blogs/news

API Documentation

SUBJCT Swagger Documentation: https://api.subjct.ai/docs/swagger/

Use Case Journey

• https://emmajaneknight.com/blogs/news/juicing-the-2025-creative-comfort-zone

1. Send the article

Pitch Deck

Certainly! Here is a concise 3-slide pitch deck summary for the hackathon idea integrating SUBJCT with MCP for content optimization:

Slide 1: Problem & Opportunity

Problem:

• E-commerce and content-heavy Shopify/WordPress sites struggle to consistently optimize internal linking, metadata, and schema at scale.
• Manual SEO enhancements are time-consuming, error-prone, and hard to maintain.
• AI-driven platforms (ChatGPT, Claude, Google ADK) need enriched, well-structured content context for better responses.

Opportunity:

• Automated content optimization + AI-context integration can boost search rankings, user engagement, and AI interactions.
• Using SUBJCT’s API combined with the emerging Model Context Protocol (MCP), we can build a scalable, automated server to transform CMS content seamlessly.

Slide 2: Solution Overview

What We Built:

• A Model Context Protocol (MCP) Server leveraging SUBJCT’s API to analyze and enrich Shopify/WordPress articles in real-time.
• Automatically injects relevant internal links, tags, schema markup, and knowledge graph data—enhancing SEO and content discoverability.
• Fully compatible with SUBJCT’s existing features and extendable to any CMS platform.

How It Works:

1. Content pulled from CMS → sent to MCP server.
2. MCP server calls SUBJCT API for content analysis & enrichment.
3. Enriched content with optimized links & metadata delivered back for publishing/AI processing.
4. MCP server plugs into LLM platforms (Claude, ChatGPT, Google ADK) for smarter AI responses.

Slide 3: Impact & Next Steps

Impact:

• Saves hours of manual SEO and content optimization work.
• Increases organic traffic and customer engagement through better site architecture.
• Enhances AI-driven tools’ effectiveness using richer, structured content context.
• Demonstrates future-proof integration between CMS, SEO automation, and AI ecosystems.

Next Steps:

• Scale MCP server to support multiple CMS integrations and features (e.g., multilingual support).
• Develop UI tools for non-technical users to customize linking rules and schema presets.
• Expand AI integration by building demo apps with ChatGPT and Google ADK using MCP.
• Explore commercializing as a Shopify App or WordPress Plugin to streamline adoption.

Install

• https://emmajaneknight.com/blogs/news/juicing-the-2025-creative-comfort-zone

    "subjct": {
      "command": "{FULL_PATH}/node",
      "args": ["{FULL_PATH}/index.ts"],
      "env": {
      }
    }