Building Agentic AI Apps with Docker MCP & Cagent
summary
Shows how to run LLMs locally with Docker Model Runner (containerized models exposing REST/gRPC endpoints, single heavy download, uses local GPUs) and how the Model Context Protocol (MCP) and Docker MCP gateway let agents call 300+ catalogued API servers with one-line config and central credential storage. Demonstrates YAML-driven agents via Docker agent/Cagent, spinning up short-lived containers per API call (GitHub queries work in demo and containers auto-remove), and publishing agent YAMLs to Docker Hub for reproducible deployment.
AI-written summary Β· about
chapters (13)
- 0:00Agentic AI and Docker MCP intro
- 1:45Gen AI vs agentic AI basics
- 2:49Docker Model Runner for local LLM
- 6:59Building agentic apps challenges
- 9:26Model Context Protocol explained
- 12:36Docker MCP gateway orchestration
- 15:27Docker MCP catalog overview
- 20:41Docker Cagent framework intro
- 24:14Agents with Docker Cagent YAML
- 27:24Multi-agent architectures
- 30:13Demo: Docker MCP in Docker Desktop
- 35:50Security in agentic workflows
- 37:14Q&A
related talks
34mPackaging & Deploying AI Tools with Model Context Protocol
43mDocker Agent β How to Make Building AI Agents Even Simpler?
30mMind the Gap - A Developer's Roadmap To Building AI Agents
25mBuilding AI Workflows: From Local Experiments To Serving Users
36mPutting AI Agents to Work: A New Era of Open Source Connectivity with MCP
40mBuilding Agents with Agent Bricks and MCP
28mAgentic AI systems using Model Context Protocol
50mCompose the Future: Building Agentic Applications, Made Simple