Agentic AI systems using Model Context Protocol
summary
Demonstrates Model Context Protocol (MCP) with Amazon Bedrock agents by wiring a Bedrock inline agent to MCP servers (HTTP transport, OAuth 2.1 auth) and community SDKs; three working demos: document QA returning Amazon facility counts, a 7‑day AWS cost pull that generates a Python bar chart in a sandbox, and a CLI QDeveloper flow that fetches live pricing to cost a 3‑tier app for 10 users/100 GB and returns a single estimated total. Covers hosting on EKS, consent fatigue and token sprawl, and untrusted third‑party MCP server risks.
AI-written summary · about
chapters (10)
- 0:00Introduction and agenda
- 2:11Understanding MCP
- 4:51The need for MCP
- 8:01MCP implementation on AWS
- 12:28Challenges and opportunities
- 16:16Security considerations
- 18:01Demo: Knowledge base retrieval
- 20:44Demo: Code interpretation
- 22:07Demo: AWS Q developer CLI
- 23:43Conclusion and resources
related talks
33mThe Model Context Protocol: Connecting AI to Everything
50mBeyond local tools: Dive into the Model Context Protocol (MCP)
46mMCP for LLM Agents, Infrastructure APIs & Data Services
26mBuilding AI-Enabled Spring Applications with Model Context Protocol (MCP)
29mEmpowering Your Product with Agentic AI
30mComposable AI: Building Next-Gen AI Agents with MCP
16mDriving agentic innovation w/ MCP as the backbone of tool-aware AI
32mMCP for dummies: Feed more context to LLM