Enterprise MCP Hackathon
Sep 13, 2025
Alchemy Studio
Alchemy Studio is an agentic AI that finds your customers, auto-generates tailored ad creatives, and runs rapid tests to see what converts, all before you launch. By functioning as an MCP (Model Context Protocol) server, it becomes an intelligent hub that allows a master AI to orchestrate its powerful internal capabilities.
Claude Monitor
Claude Monitor gives managers a way to track their employees AI usage. It rewrites their crappy prompts using the best practices of prompt engineering, shows them side-by-side comparisons of prompts they wrote vs how AI rewrote them, allows them to generate custom prompt engineering improvement reports on a per-conversation-basis, and bonus: auto-updates their CLAUDE.md with errors that the AI made in its convo, so that they AI never makes the same error twice.
Fugu
Fugu is a security scanner that automatically audits MCP servers for vulnerabilities before you deploy them. Paste any GitHub repo or MCP server URL, get an instant security report identifying attack vectors like tool poisoning, code injection, and data exfiltration risks.
gullie
fullstack AI relocation platform - ai agents helping companies relocate employees across 160+ countries
Markdown MCP
Markdown MCP is a more accurate and token efficient framework for building MCP servers. Instead of dumping dozens of tool calls onto a user's agent or 1000s of tokens into their context window from an llms.txt... Markdown MCP is faster and more token efficient way to use MCP by using just 3 tool calls.
MCP-B/WebMCP
> NOTE: our apps our now deployed. We did not see the submission criteria at the time of recording. Convex react workos app: https://didactic-octo-parakeet.vercel.app/ We implemented and expanded on the declarative webMCP api first proposed here: https://github.com/webmachinelearning/webmcp/issues/22 Apps just need to add some html tags to their html and the extension will convert them into webMCP tools which it can then call. these tags can also been added by the extension via scripting
Pink Dot
I work for an enterprise company that wants to encourage AI and MCP adoption across the org. But we have a huge problem: Security. Enterprise companies don't have a good way to specify which tools their employees are allowed to use. MCP is currently the wild west. Employees can install and connect to any MCP they find on the internet. I want to fix that - watch the video to see how! Site: https://mcp-rbac.vercel.app MCP: https://mcp-rbac.vercel.app/api/mcp
playwright-orchestrator
Problem I wan to solve is to be able to have multiple claude sessions run a playwright mcp without it breaking. Currently a playwright mcp uses the same chromium so it doesn't allow parrallel claude sessions. mcp-playwright-orchestrator creates a docker instance of a playwright-mcp and manages allowing a claude session to have it's own docker instance of a playwright mcp leading to a huge opportunity of lead generation and workflow automation in my company. For example searching for ecommerce stores, cafe stores for square, or automating pos ordering for square.
Refactor Project
Solves three-way merges using AI based on a set of defined program requirements, and provides a useful MCP interface to systematically refactor projects with interlinked modules.
Second Self
Second Self ensures sales reps only engage leads ready to buy. Our AI agents automatically nurture prospects across phone, email, LinkedIn, Slack, iMessage, IG, and WhatsApp, scoring engagement and intent. When leads enter your pipeline, agents evaluate readiness using internal Attio records and external Exa data to enrich. High-intent prospects go straight to sales, while others receive personalized AI follow-up. As agents track growing interest signals, they instantly hand off leads the moment they cross your readiness threshold. Your team stays focused on closing deals with buyers ready to move forward, not chasing cold leads. Stop wasting time on unqualified prospects; let AI warm them until they're ready to convert.
