Back to Products

Mrinanka's Tools MCP Server - AI-Native Portfolio

A production MCP (Model Context Protocol) server that transforms my portfolio into an interactive AI data source. Anyone using an AI client, from Claude Desktop to terminal tools, can instantly explore my work, ask questions, and get in touch directly through their preferred interface.

Mrinanka's Tools MCP Server - AI-Native Portfolio
Live Demo

Technologies Used

MCPStreamable HTTPResendRAGTypescript

The MCP (Model Context Protocol) Server is an AI-native interface built directly into my Next.js portfolio. It allows anyone using an AI client or coding assistant to programmatically interact with my professional presence, instantly access my experience, and get in touch without ever leaving their terminal or app.

The Problem

Traditional portfolios are static. While human visitors can read about you, AI agents cannot programmatically interact with your site. As AI-native interfaces become the primary way people discover and evaluate work, portfolios need to speak the language of machines. I wanted to make learning about my background as effortless as asking an AI a question.

Key Features

  • Universal Access: Connect directly from Claude Desktop, Cursor, Windsurf, or terminal-based AI tools.
  • getKnowledgeBase: A tool that serves a curated Markdown version of my professional profile, allowing AIs to instantly answer questions about my experience.
  • sendMessageToMrinanka: A secure tool that collects visitor inquiries and delivers them directly via the Resend API.
  • XSS-safe HTML escaping for all email templates
  • Bearer token API key verification for production security

Architecture

Built as a serverless-first Next.js App Router API route (/api/mcp), it uses Streamable HTTP transport via the @modelcontextprotocol/sdk. This means zero cold-start overhead and no extra infrastructure to manage. Every tool input is strictly validated at the protocol level using Zod, ensuring complete type safety.

Why MCP?

MCP is an open standard that provides a universal protocol for AI agents to discover and invoke tools. By adopting this standard, the portfolio becomes effortlessly discoverable. Whether it's a recruiter using Claude Desktop or a developer in their terminal, anyone can interact with my portfolio, query information, or send messages directly from their preferred AI environment without any custom integration.