A few Agent Development Kit (ADK) Articles to dive into : June 1–10, 2025
The Agent Development Kit (ADK) is seeing solid traction with articles covering it across its various capabilities.
The Agent Development Kit (ADK) is seeing solid traction with articles covering it across its various capabilities. Check out key articles on Google Cloud Medium in this category in the last 10 days.
✅ Shir Meir Lador has started a series titled “A Beginner’s Guide to Building an AI Trend Finder with ADK“. The first part (https://bit.ly/4e1uTic) covers building your first agent using ADK and then the next part (https://bit.ly/43EdHfe) dives into custom tools, orchestrator, debugging and more. If you are looking to not just get started with ADK but also quickly move on to building on a complex real-world agent, bookmark this series.
✅ As of Agent Development Kit (ADK) v1.2.0+ you can deploy your ADK agents directly to Vertex AI Agent Engine without any additional code! Yes, you read that correctly. Just a single command `adk deploy agent_engine ….` to do that. Check out Jack Wotherspoon’s guide (https://bit.ly/3Th3KOD) to doing this.
✅ ADK v1.2.0+ has made the process of integrating MCP Tools into your Agents much cleaner. Arjun Prabhulal demonstrates (https://bit.ly/3Tkv8v8) how you can do that via “Building ADK(v1.2.0+) agents using MCP tools”.
✅ Guillaume Laforge wants to ensure that Java programmers have access to a wider variety of models than what the ADK Java release currently supports (Gemini & Claude). Guillaume has approached this challenge by integrating Langchain4j and ADK together, so that any model that Langchain4j supports can be made available to ADK too. Read it here: https://bit.ly/43X5Jwz
✅ The MCP Toolbox for Databases is a great way to add tools to your existing environments that support MCP & to work efficiently with a wide range of databases, including several Google Cloud databases. You could integrate this toolbox capability inside your Agents too. Billy Jacobson discusses this in his article “Beyond the Buzz: Practical AI with Cloud Databases using MCP and ADK” (https://bit.ly/3SOKeZL)
✅ When an Agent executes your code, there are several hooks/callbacks that it provides for you to tap into the execution. These include before_, after_ for each of the following: agent, model and tool callbacks. These play a crucial role when it comes to building out safeguards. Glen Yu covers the scenarios (https://bit.ly/4kYpVou) with a few practical examples.
✅ What is Spicy Black Garlic Chicken McNuggets from Japan, Pistachio McFlurry from Italy and Cheese & Bacon Loaded Fries from Australia. These are the items on the weekly international menu from McDonalds. Adrien SIEG argues “Why click buttons when agents can do it for you?” and demonstrates a proof of concept on what you need to do to build a “multi-agent system that automatically orders my weekly McDonald’s fix using Google’s ADK, Selenium and more”. (https://bit.ly/4jKwx90)
If you would like to write about your experience with the Agent Development Kit, do fill out this form to register as a writer on Google Cloud Medium publication. We’d love to publish and share your expertise with the community.
