Docker Agent — How to Make Building AI Agents Even Simpler?
summary
Introduces Docker Agent, an open-source toolkit to build ad-hoc AI agents without coding by assembling models, prompts and tools; it works with any LLM (local or remote), any tool interface, and connects via HTTP, A2A, MCP or ACP. Shows how Docker Agent replaces large Python codebases (example: Docker’s Gordon v2 implemented from many Python files into one YAML) and teaches attendees to create single agents and fleets for reusable or disposable tasks like coding, cooking or writing.
AI-written summary · about
chapters (7)
- 0:00Docker Agent intro
- 8:51Demo: useless agent
- 13:01Code agent with Docker Agent
- 24:10Business agents
- 31:30Agent testing & optimization
- 38:11Integrating agents
- 42:45Q&A
related talks
39mBuilding Agentic AI Apps with Docker MCP & Cagent
30mMind the Gap - A Developer's Roadmap To Building AI Agents
31mNavigating Containers: AI Powered by Docker
29mDemystifying AI Agents with Python Code
07h 25mAI Agents Conference 2026 - Day 2
25mBuilding AI Workflows: From Local Experiments To Serving Users
24mDevOps with AI Agents in Practice
01h 21mShip Agents that Ship: A Hands-On Workshop