Agent Development Kit (ADK)

What is Agent Development Kit (ADK)?

Agent Development Kit (ADK) is an open-source Python framework from Google for building, evaluating and deploying AI agents. You define agents in code with the Agent class, then arrange them with Workflow, a graph-based engine that supports routing, fan-out and fan-in, loops, retries, state, human-in-the-loop steps and nested workflows. A Task API handles agent-to-agent delegation, and tools can be pre-built, custom functions, OpenAPI specs or MCP tools. It is tuned for Gemini but model-agnostic and deployment-agnostic. Install it with pip install google-adk on Python 3.11 or newer, test locally with the adk run command or the adk web development UI, run bundled evaluations with adk eval, and deploy to Docker, Cloud Run or Vertex AI Agent Engine. Versions exist for Java, Kotlin, Go and TypeScript, and an Agent Config feature builds agents without code. ADK is free under the Apache 2.0 license.

Use cases

• Build a Python AI agent with tools
• Orchestrate several specialized agents
• Create graph-based agent workflows
• Evaluate agent behavior with test sets
• Add human approval before tool calls
• Deploy agents to Cloud Run
• Debug agents in a local web UI

Who it's for

• Python developers
• AI engineers
• Platform teams
• Researchers

Why use Agent Development Kit (ADK)

• Agent logic, tools and orchestration live in ordinary Python, so they can be tested and versioned like other code.
• The Workflow engine gives deterministic control over how multiple agents hand work to each other.
• A local web UI and an eval command help you debug and score agents before shipping.
• You are not locked to one model or host; Docker, Cloud Run and Vertex AI Agent Engine are all supported deployment paths.
• It costs nothing to use under the Apache 2.0 license, with community integrations in a separate repo.

Agent Development Kit (ADK) features

• Graph-based Workflow runtime with routing, loops, retry and nested workflows
• Agent class defining instructions, tools and model
• Task API for structured agent-to-agent delegation
• Multi-agent hierarchies of specialized agents
• Tools from pre-built sets, functions, OpenAPI specs and MCP
• Tool confirmation flow for human-in-the-loop checks
• Built-in development UI via the adk web command
• Evaluation through the adk eval command
• Deployment to Docker, Cloud Run and Vertex AI Agent Engine
• Agent Config for building agents without code

Agent Development Kit (ADK) integrations

• Gemini
• Vertex AI Agent Engine
• Cloud Run
• Docker
• MCP

How to get started with Agent Development Kit (ADK)

1. Install the package with pip install google-adk on Python 3.11 or newer.
2. Define an Agent with a name, model and instruction, and attach tools such as functions, OpenAPI specs or MCP tools.
3. Compose agents into a Workflow with edges to control the order, routing and loops.
4. Run it with adk run, or open adk web to test and debug in the development UI, then score it with adk eval.
5. Deploy with adk deploy docker or adk deploy cloud_run, or scale on Vertex AI Agent Engine.

Agent Development Kit (ADK) limitations

• Requires Python 3.11 or newer and some programming
• Optimized for Gemini and Google Cloud deployment, though other models and hosts are supported
• The main branch can contain experimental changes not yet in a stable release

Agent Development Kit (ADK) pricing

Agent Development Kit (ADK) is open source, so you can use and self-host it for free.

Prices can change and may vary by country or exclude tax. Check the official pricing page.

Agent Development Kit (ADK) alternatives

The main Agent Development Kit (ADK) alternatives listed on Debuly are LangGraph, OpenAI Agents SDK, CrewAI, Microsoft Agent Framework, Strands Agents and AutoGen. The table compares pricing and platforms; order is A–Z.

1. 

   AutoGen

   Microsoft's MIT-licensed Python and .NET framework for multi-agent AI apps, now archived; new projects are pointed to Microsoft Agent Framework.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit AutoGen (official site)
2. 

   CrewAI

   Multi-agent framework and platform for building AI agent crews in Python or a visual editor, then deploying and monitoring them.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit CrewAI (official site)
3. 

   LangGraph

   Low-level open-source orchestration framework and runtime for building stateful, long-running AI agents as graphs.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit LangGraph (official site)
4. 

   Microsoft Agent Framework

   Open, multi-language framework from Microsoft for building, orchestrating and running AI agents and multi-agent workflows in Python, .NET and Go.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit Microsoft Agent Framework (official site)
5. 

   OpenAI Agents SDK

   Open-source Python framework for building multi-agent workflows with tools, handoffs, guardrails, tracing, sandboxes and voice agents.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit OpenAI Agents SDK (official site)
6. 

   Strands Agents

   Free Apache-2.0 SDK for creating and running Python and TypeScript AI agents, with built-in tools, MCP, memory and multi-agent patterns.

   Developer platformsAutomation and agentsCoding
   Open sourceVisit Strands Agents (official site)

Agent Development Kit (ADK) FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from avatars.githubusercontent.com · fetched 24 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.