Letting AI Interface with Your App with MCPs
summary
Argues that a standard Model Context Protocol replaces bespoke integrations by having services expose an MCP server so any AI client can orchestrate them, showing a live demo where an AI assistant authenticates via email token, calls EpicMe and locationator and a weather MCP, and ultimately called 16 tools to create a journal entry. Shows MCP UI inside VS Code, cites Stripe MCP for billing, and claims this inversion of control makes multi-service tasks seamless and actually work today.
AI-written summary · about
chapters (12)
- 0:00Introduction and warm-up
- 2:44What is Jarvis?
- 5:37AI's paradigm shift in software
- 8:08Why don't we have Jarvis yet?
- 10:36Introducing MCP
- 14:50MCP's inversion of control
- 16:14MCP architecture explained
- 17:39Live MCP demo
- 25:24MCP UI and user experience
- 27:08MCP in consumer tools
- 29:59Learning and building with MCP
- 31:02Closing and Q&A
related talks
20mLetting AI Interface with your App with MCP
20mLetting AI Interface with Your App with MCPs
30mComposable AI: Building Next-Gen AI Agents with MCP
35mMCP in Action: Real-World Use Cases & Design Patterns
33mThe Model Context Protocol: Connecting AI to Everything
16mOne AI Chat to Rule Them All: Connecting Enterprise Tools with MCP
21mIntegrate Your APIs into the New AI Marketplace
29mEmpowering Your Product with Agentic AI