Command Palette

Search for a command to run...

Google ADK Samples

Active

A collection of sample agents built with Agent Development (ADK).

Ken's Contributions

  • Travel Concierge Optimization: Resolved critical type declaration inconsistencies within the Travel Concierge sample agent, ensuring strict schema validation for the Gemini-powered reasoning engine (#32).

More contributions on GitHub.

Overview

The Agent Development Kit (ADK) is a flexible and modular framework for developing and deploying AI agents. While optimized for Gemini and the Google ecosystem, ADK is model-agnostic and built for compatibility with various orchestration frameworks.

This repository serves as the definitive collection of Sample Agents—functional starting points for common application scenarios like retail assistance, travel planning, and complex multi-agent workflows.

Features

  • Ready-to-Use Logic: Pre-packaged core logic for diverse use cases (Retail, Travel, etc.) including tool usage and evaluation strategies.
  • Human-in-the-Loop: Samples demonstrating how to integrate human oversight into agentic workflows for improved reliability.
  • Model Agnostic: While highly optimized for Google's Gemini, the samples are designed to work across different LLM backends.
  • Developer-Centric: Structured to make agent development feel more like traditional software development (testing, versioning, orchestrating).

Architecture & Tech Stack

ADK is built on a "model-agnostic, deployment-agnostic" philosophy, allowing developers to orchestrate agentic architectures ranging from simple tasks to complex, autonomous workflows.

  • Frontend/Client: Samples include TypeScript and JavaScript integrations for web-based agent interfaces.
  • Backend/Logic: Primarily built with Python, utilizing its robust AI ecosystem and the foundational ADK library.
  • Data & Evaluation: Leverages Jupyter Notebooks for evaluation and data processing, with Go used for high-performance utility components.

Milestones & Impact

Documenting the evolution of ADK from a foundational framework to a rich ecosystem of agent samples.

2025: Ecosystem Expansion

  • Strategic Samples: Introduced a wide range of industry-specific agents (Travel, Retail, Customer Support) to accelerate developer onboarding.
  • Multi-Agent Orchestration: Deployed advanced samples demonstrating complex agent-to-agent communication and task delegation.

2024: Framework Foundation

  • ADK Core Development: Initial development and launch of the foundational framework, establishing the core technical stack and design principles.
  • Documentation Launch: Established the initial documentation ecosystem for the framework.