adk-python

An open-source, code-first Python toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control.

What is adk-python?

An open-source, code-first Python toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control.

Overview of adk-python

What it is

\n\nAgent Development Kit (ADK) is an open-source, code-first Python framework for building, evaluating, and deploying sophisticated AI agents. It applies software development principles to AI agent creation, designed to simplify building, deploying, and orchestrating agent workflows from simple tasks to complex systems. While optimized for Gemini, ADK is model-agnostic, deployment-agnostic, and compatible with other frameworks.\n\n

Key features

\n\n

  • Workflow Runtime: A graph-based execution engine for composing deterministic execution flows with support for routing, fan-out/fan-in, loops, retry, state management, dynamic nodes, human-in-the-loop, and nested workflows.\n
  • Task API: Structured agent-to-agent delegation with multi-turn task mode, single-turn controlled output, mixed delegation patterns, human-in-the-loop, and task agents as workflow nodes.\n
  • Modular Multi-Agent Systems: Design scalable applications by composing multiple specialized agents into flexible hierarchies.\n
  • Rich Tool Ecosystem: Utilize pre-built tools, custom functions, OpenAPI specs, MCP tools or integrate existing tools for tight integration with the Google ecosystem.\n
  • Code-First Development: Define agent logic, tools, and orchestration directly in Python for flexibility, testability, and versioning.\n
  • Agent Config: Build agents without code using the Agent Config feature.\n
  • Tool Confirmation: A tool confirmation flow (HITL) that guards tool execution with explicit confirmation and custom input.\n
  • Deploy Anywhere: Easily containerize and deploy agents on Cloud Run or scale with Vertex AI Agent Engine.\n\n

Use cases

\n\n

  • Building and deploying AI agents for various tasks, from simple assistants to complex multi-agent systems.\n
  • Evaluating agent performance with built-in evaluation tools.\n
  • Orchestrating agent workflows with deterministic execution flows, including human-in-the-loop interactions.\n
  • Integrating with the Google ecosystem and deploying on Google Cloud infrastructure.\n\n

Getting started

\n\nInstall the stable release:\n\nbash\npip install google-adk\n\n\nRequirements: Python 3.11+.\n\nQuick start example (Agent):\n\npython\nfrom google.adk import Agent\n\nroot_agent = Agent(\n name=\"greeting_agent\",\n model=\"gemini-3.5-flash\",\n instruction=\"You are a helpful assistant. Greet the user warmly.\",\n)\n\n\nQuick start example (Workflow):\n\npython\nfrom google.adk import Agent, Workflow\n\ngenerate_fruit_agent = Agent(\n name=\"generate_fruit_agent\",\n instruction=\"Return the name of a random fruit. Return only the name.\",\n)\n\ngenerate_benefit_agent = Agent(\n name=\"generate_benefit_agent\",\n instruction=\"Tell me a health benefit about the specified fruit.\",\n)\n\nroot_agent = Workflow(\n name=\"root_agent\",\n edges=[(\"START\", generate_fruit_agent, generate_benefit_agent)],\n)\n\n\nRun locally:\n\nbash\nadk run path/to/my_agent\nadk web path/to/agents_dir\n\n\n

License and pricing

\n\nLicensed under the Apache 2.0 License. Open source and free to use.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/googleadk-python.svg)](https://yunodb.com/tools/googleadk-python?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/googleadk-python?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/googleadk-python.svg" alt="Featured on YunoDB" height="20"></a>