MCP ExplorerExplorer

World Wild Web Ai And Mcp Hack Night

@Nleaon 5 days ago
3 MIT
FreeCommunity
AI Systems
a repo to describe the hack challenge for the event

Overview

What is World Wild Web Ai And Mcp Hack Night

World-Wild-Web-AI-And-MCP-Hack-Night is an event focused on developing AI agents and backend services using MCP servers. Participants explore their creativity by building applications that leverage artificial intelligence and real-world integrations.

Use cases

Potential use cases for the hacknight include creating personalized AI assistants such as language learning tutors, mental wellness coaches, email assistants, music discovery tools, smart home orchestrators, and personal customer relationship management (CRM) systems.

How to use

To participate, attendees must clone the provided template repository, follow the setup instructions, and test their AI agent locally. During the hacknight, participants will access Codegen for MCP server creation, connect their agents, and start building features.

Key features

Key features of the project include a built-in connection to MCP servers, integration with Fiberplane’s AI agents playground, and access to Anthropic credentials for LLM. Additionally, there are resources for generating MCP servers and Hono APIs.

Where to use

Projects can run on Cloudflare infrastructure, allowing for scalability and real-world applications or can be developed locally by participants as part of the hacknight challenge.

Content

World-Wild-Web-AI-And-MCP-Hack-Night

Hey, thanks for joining us for WWW AI & MCP Hacknight! In this repo, you’ll find all the information you need to get started with the hacknight.

We recommend getting familiar with the challenge before the hacknight, as it will help you get started faster.

Prerequisites

  • Your own laptop
  • Node.js and a JS package manager (npm, yarn, pnpm, bun) installed
  • A Cloudflare account
  • A GitHub account

The Challenge

This hacknight is all about exploration and creativity with AI agents and MCP servers! Build something that excites you - whether it’s practical, experimental, or just fun.

Core Components

Your project will consist of two main parts:

  1. An AI Agent: A chat-based assistant with its own personality and purpose
  2. Backend Services: MCP server, agent tools and APIs that give your agent real-world capabilities

What We Provide to Get Started:

  • A template based on the Cloudflare Chat Agent Starter Kit with:
  • Access to Fiberplane’s Codegen platform for generating:
    • MCP servers that run on Cloudflare
    • Hono APIs for integrating with external services (Notion, GitHub, etc.)
  • Anthropic credentials for the LLM that controls the agent. Credentials will be provided for the duration of the hacknight.

Quick Start

  1. Clone the template from the ai-agent-template directory
  2. Follow the setup instructions in the template’s README
  3. Test that your AI agent works locally

During the Hacknight:

  1. Get access to Codegen and create your MCP server
  2. Connect your agent to the MCP server (Note: Choose an MCP server without authentication)
  3. Start building your features!

Need Inspiration?

While you’re free to build anything, here are some ideas for your themed AI agent. Think of the overall topic: Personal AI Automation

  • A language learning tutor
  • A mental wellness coach
  • An email assistant
  • A music discovery assistant
  • Smart Home orchestrator
  • Personal CRM

Fiberplane Codegen Project & Anthropic API Key

For access to the early pre-beta version of Fiberplane’s Codegen project, please join Fiberplane’s Discord and share your GitHub handle in the #ldn-hacknight channel. We will add you to the platform and also share the Anthropic API key for the Hack Night.

Examples, Resources and Inspiration

Evaluation Criteria

  • Extra points for shared GitHub repo
  • Live demos are preferred over presentation slides
  • Agents and MCP server should run on Cloudflare (can be also locally)
  • Extra points if you add a code-generated MCP server to your AI agent
  • Nothing to demo? Don’t worry - you can share your code and learnings from the evening

Submission

If you’d like to present your project:

  1. Fill out this form
  2. Join the Google Meet call we’ll share near the end of the event

We’re using a shared call to make presentations smooth and avoid switching between computers.

Tools

No tools

Comments