Skip to Main Content
Talk Beginner First Talk

Building MCP Servers: Connecting AI Agents to the Real World

Proposal status is Review Pending
Session Description

AI agents are great at thinking, but pretty useless if they can’t connect to the tools we use every day. MCP (Model Context Protocol) is a standard that changes that, it lets agents talk to APIs, databases, and pretty much anything using API's.

Key Takeaways

Key Takeaways

  1. What MCP actually is – and why it’s worth knowing if you want AI agents to do more than just chat.

  2. How MCP servers work – the architecture, request/response flow, and integration points.

  3. Steps to build your own MCP server – from setting up the basics to adding custom tools.

  4. Making it production-ready – arguments, environment variables, multiple transport options (SSE & HTTP streaming).

  5. Applying it anywhere – how the same approach can connect AI agents to any system, not just the demo tool.

References

https://youtu.be/GNQcrFcRQys?feature=shared
https://github.com/CW-Codewalnut/metabase-mcp-server

Session Categories

Other
Open source MCP server

Speakers

Ranjan Bhagat
Ranjan Bhagat Senior Software Engineer | Codewalnut

Hi, I am Ranjan Bhagat is a passionate software engineer with over 4 years of hands-on experience building applications across diverse domains using Flutter and React.js. Alongside crafting mobile and web apps, I have a keen interest in AI-driven systems. I recently created the Metabase MCP Server, an open-source project that bridges AI agents with real-world tools (Metabase, which is an BI tool like Power BI and Looker)

https://www.linkedin.com/in/ranjan-bhagat-326558215