ADK Masterclass — Map of Content
Hub for everything learned about Google’s Agent Development Kit (ADK). Source material: a masterclass / internship-era set of notes I took while building agents on Vertex AI.
Start here
- ADK-Overview — what ADK is, why it exists, key features
- ADK-Basics-and-Setup — virtualenv, agent properties, folder structure, CLI
Core building blocks
- ADK-Tools — function tools, built-in tools, third-party tools
- ADK-Connecting-Other-LLMs — LiteLLM + OpenRouter for non-Gemini models
- ADK-Structured-Output —
input_schema,output_schema,output_key - ADK-Sessions-and-State — session types, state, runners
- ADK-Database-Sessions — persistent sessions via SQLite / cloud
- ADK-Callbacks — the 6 callback types (before/after agent, model, tool)
Multi-agent patterns
- ADK-Multi-Agent — delegation, AgentTool vs sub-agent
- ADK-Stateful-Multi-Agent — combining state with multi-agent
Workflow agents
- ADK-Sequential-Agent — run sub-agents in order, share via state
- ADK-Parallel-Agents — run sub-agents concurrently
- ADK-Loop-Agents — iterate until termination condition
Related (outside this MOC)
- RAG-Overview — RAG concepts (the RAG agent built in this course uses ADK)
- Vertex-AI-RAG-Agent — full ADK + RAG agent example