Unleashing MCP: Building Remote AI Capabilities Without the Integration Hell
Session Description
Modern AI workflows are shackled by static integrations—every new tool or API requires manual code, slowing innovation. MCP (as I like to call it, Missing Capability Protocol) breaks these chains. Unlike conventional API wrappers, MCP enables AI agents to dynamically discover, authenticate, and use any capability at runtime without pre-defined integrations.
In this lightning talk, we'll build a live MCP server and demonstrate how it allows AI agents to autonomously:
Seamlessly connect to unknown services (e.g., calendars, APIs, databases).
Compose multi-step workflows using never before seen tools.
Eliminate brittle, hardcoded integrations from the AI development loop.
See firsthand how MCP transforms AI from a path follower into an adaptable problem-solver
Key Takeaways
Breaking static AI workflows: How MCP eliminates brittle, hardcoded integrations that slow AI innovation
Dynamic capability discovery: The paradigm shift from pre-built integrations to runtime service discovery and authentication
Autonomous AI agents: How to build AI systems that can compose multi-step workflows with never-before-seen tools
Hands-on MCP implementation: Step-by-step process to create remote MCP servers that transform AI from script-followers into adaptable problem-solvers
Session Categories
Which track are you applying for?
Speakers
Pratik
Reviews
There are 7 proposals on MCP. I'd mark this in top2 among them- with less talk and more tutorial. The proposer already has done this talk in less than 10 minutes, so it would work as a Lightning talk.
This seems quick and informative. People should know about MCP, and as an emerging technology we still dont knowall the implications but honing in on some practical ones is what people need.
The proposal is detailed, but there has already been a significant amount of public discourse on how to create MCP servers, so I'm not sure about this proposal.
Thank you for submitting your proposal for IndiaFOSS 2025. Your submission was well-received and progressed to our final review stages.
Unfortunately, due to the high volume of excellent proposals this year, we were unable to select your talk for the final program. We appreciate the effort you put into your submission and encourage you to apply again for future events.