Create Multi-Agent Systems with ADK
summary
A live demo shows how to build multi-agent systems using the ADK (Agent Development Kit) by running a Python ADK repository in Google Cloud Shell and using Gemini 2.5 Flash as the LLM via ADK’s CLI and web UI. The setup composes a parent router agent and multiple subagents that route requests, execute sequential pipelines, loop until acceptance, and run parallel queries, plus a custom tool that writes selected items into ADK state (saved attractions), demonstrating coordinated agent responses and persisted state.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 5:19What is ADK and its advantages
- 12:24Agent patterns: router, sequential, loop
- 21:57Parallel agents and use cases
- 33:24Lab setup
- 47:00Tools and state in ADK
- 51:14Demo and conclusions
related talks
01h 09mMulti Agent Architecture using ADK
45mUnifying AI experience for customers: Building multi-agent systems with Google ADK
57mCoding with ADK to beat a text adventure game
01h 03mSupercharge your AI agents: Inside the new ADK integrations ecosystem
30mLive Demo – Develop and Deploy a Trustworthy Multi-Agent System
49mLet's build an AI agent
34mBuilding Multi-Agent Generative-AI Applications with AutoGen
01h 04mAgentic AI: Build a Multi-Agent Application with CrewAI