Global MCP Hackathon

Use Case

As autonomous agents move from research to production, they need a reliable interface layer—secure, permissioned, and observable—to interact with real-world APIs. Most existing APIs were not designed with agents in mind, lacking the identity, scoping, and trust frameworks required to operate autonomously at scale.

This challenge asks you to close that gap.

Your MCP server will act as a building block in this new generation of software architecture: one where agents can safely invoke tools, personalize experiences, and automate operations—with authentication with natively integrated Descope and connection visibility controlled by Cequence.

By participating, you’re helping to define the standards and best practices for how AI agents securely interact with your applications and APIs.

Focus Areas

Explore these example tool ideas to spark your creativity, or feel free to come up with your own — whether it’s solving a real-world problem, streamlining a workflow, or just building something fun and experimental.

1. Agentic Shopping Experience: Explore & Buy Online

• Description: An agentic shopping experience leverages AI agents to create a dynamic, personalized, and autonomous journey for shoppers on your e-commerce site. Unlike traditional online shopping, agentic commerce uses intelligent agents that anticipate needs, provide real-time recommendations, and can even complete purchases on the shopper’s behalf—while keeping the user in control

• Example Features:

• Conversational Product Discovery

• Intelligent Comparison

• Seamless Purchase & Checkout

2. AI-Powered Workplace Search

• Description: Build an MCP server that lets agents search, retrieve, and summarize content from tools like Google Drive, Notion, Slack, or Confluence. Agents can help users find documents, answer questions, or pull recent updates while respecting authentication and access scopes. This service gives agents structured, secure access to internal knowledge.

• Example Features:

• Search documents, messages, and knowledge base entries

• Summarize or extract key points from retrieved content

• Support scoped, user-specific access with Descope

• Secure all agent interactions using the Cequence MCP proxy
  
  

3. Personalized Content Recommendation Engine

• Description: A backend service that provides personalized content recommendations based on user preferences, activity, and browsing history. The API can be integrated with websites, apps, or smart devices.

• Example Features:

• Recommend articles, videos, or music based on user behavior.

• Personalized shopping suggestions based on previous purchases.

• Dynamic recommendations for fitness or wellness routines based on goals.
  
  

4. Virtual Health Assistant

• Description: A tool that assists users in managing their health and wellness by integrating with wearable devices, collecting health metrics, and providing actionable insights.

• Example Features:

• Track steps, calories, and heart rate via wearable device integration.

• Suggest personalized workouts or stretches based on user health data.

• Provide daily reminders for medication or hydration.
  
  

5. Legal Document Analysis Tool

• Description: A tool that processes legal documents and extracts key clauses, terms, and conditions. It can provide high-level summaries or answer questions based on the content of the document.

• Example Features:

• Extract specific clauses (e.g., payment terms, confidentiality).

• Summarize legal language for easier understanding.

• Answer questions like "What are the terms for cancellation?" or "What is the length of the contract?"
  
  

6. AI-Powered Job Matching

• Description: An MCP server that analyzes resumes, job descriptions, and candidate profiles to match individuals with the most suitable job opportunities.

• Example Features:

• Match candidates with job roles based on skills, experience, and preferences.

• Generate personalized cover letters or interview prep based on job requirements.

• Offer career advice based on trends in the job market.

These focus areas are here to inspire you, but you are not limited to them. If you have a unique idea, we encourage you to pursue it and get creative.

Tech Stack

While you're free to choose your tools and frameworks, here are some recommended components:

• Required Tech: Implementing the MCP Server using Cequence AI Gateway 

• Hosting Platforms: Fly.io, Render, Vercel, or any preferred Jamstack hosting platform

• Agent Clients: Claude Desktop, Cursor, Windsurf, Cloudflare MCP Playground, custom MCP client

• Languages & Frameworks: Node.js or Python (FastAPI, Flask)
  
  

Guidelines

Follow these guidelines to ensure your submission meets the challenge requirements:

• Host at least one working MCP-compatible tool that agents can invoke via your server

• Implement the MCP Server using Cequence AI Gateway (Required)

• Expose your MCP Server over the public internet via a secure HTTPS endpoint

• Document your API endpoints and required scopes so agents know how to interact with your tools

• Test your service with at least one MCP client, such as Claude Desktop or Cloudflare MCP Playground, to confirm compatibility