300+ Tools CoveredSource Data Updated Weeklydates

Tool intelligence profile

Google Agent Development Kit

Google's open agent framework — hierarchical multi-agent composition, graph workflows that mix deterministic code with model reasoning, a local dev UI and evaluation harness, and a deployment path onto Vertex AI Agent Engine. Python, TypeScript, Go, Java and Kotlin.

Visit Site →
Type
Agent Framework
Category
Pricing
Free (open source)
Deployment
Cloud or self-hosted
Best for
Enterprise / polyglot teams, especially on Google Cloud
Last updatedSeptember 20, 2026Open Source

Editor's Take

Best for enterprise teams whose agent has to live inside an existing JVM, Go, or Kotlin service, or who need a partly deterministic workflow expressed as a graph rather than forced through a model — five first-party language implementations and an evaluation harness are both genuinely rare. The trade-off is gravity: the framework is portable but the deployment story that makes it attractive runs onto Vertex AI Agent Engine. Not a fit for a small Python team shipping one agent, where PydanticAI or the OpenAI Agents SDK gets there with far fewer concepts.

— Egor Burlakov, Editor

Evaluate Google Agent Development Kit

Comparisons

Google Agent Development Kit: product and architecture

This Google Agent Development Kit review looks at ADK, the open agent framework Google publishes for building, evaluating, and deploying agents — and at the question that decides whether it is right for you: how much of your stack is already Google's. ADK is Apache-2.0 and free, with implementations in Python, TypeScript, Go, Java, and Kotlin, and a deployment path onto Vertex AI Agent Engine. We evaluated it against its documentation, its GitHub repositories, and the adoption signals we collect weekly.

Overview

ADK is Apache-2.0 licensed and published by Google. Our 14 September 2026 snapshot recorded about 21,500 GitHub stars on google/adk-python and roughly 2.4 million PyPI downloads a week for the google-adk package. A separate google/adk-java repository carries the Java implementation, and the project also ships TypeScript, Go, and Kotlin.

Its market position is the enterprise, polyglot end of the agent market. The project's own framing is "build production agents, not prototypes", and the five-language surface is the clearest expression of who it is for: organisations where the agent has to live inside an existing JVM or Go service rather than in a Python notebook. Version 2.0 introduced graph workflows, and the Python package was at v2.9.2 in mid-September 2026.

The audience is enterprise engineering teams, particularly those already on Google Cloud, who need agents that deploy into managed infrastructure and pass an evaluation gate before they do.

Key Features and Architecture

The base unit is an Agent with a name, a model, an instruction, and a list of tools — the same four fields in every language. From there, ADK's distinguishing features are structural rather than ergonomic.

Graph workflows, new in ADK 2.0, are the headline. They let you weave deterministic code with model reasoning in one structure: explicit execution paths where the process is known, model reasoning where it is not. This is the feature that separates ADK from frameworks that offer only a loop, and it is the reason a compliance-sensitive workflow can be expressed here without pretending the whole thing is deterministic.

Hierarchical multi-agent composition lets agents contain agents, so a coordinator delegates to specialists and the structure is a tree you can read rather than a set of handoff rules you have to trace.

The Agents CLI scaffolds, builds, tests, evaluates, and deploys from the terminal, and is designed to be driven from an AI-enabled editor.

Evaluation is first-party rather than an afterthought. ADK ships a harness for scoring agent behaviour, which matters because the thing that stops an agent reaching production is usually the absence of a pass/fail gate, not the absence of a feature.

Tooling includes Google's own — google_search is a one-line import — alongside an open ecosystem for connecting other services.

Deployment targets Vertex AI Agent Engine as a managed runtime, and this is where the Google gravity is strongest. ADK runs anywhere; the smooth path runs onto Google Cloud.

Ideal Use Cases

ADK earns its place in four situations. The first is a polyglot enterprise. If the agent has to be a Java service because the surrounding system is Java, ADK is one of very few frameworks that does not force a Python sidecar, and Go and Kotlin get the same treatment.

The second is a workflow that is partly deterministic. Graph workflows are the right shape when some steps must happen in a fixed order with explicit branches — an approval, a policy check, a write to a system of record — and only some steps need a model to decide.

The third is a team on Google Cloud that wants a managed runtime. Deploying to Vertex AI Agent Engine removes the operational question entirely, and the integration is first-party rather than reconstructed.

The fourth is any team that has to demonstrate an agent works before shipping it. The built-in evaluation harness is the mechanism, and adding one to a framework that lacks it is real work.

Do not choose ADK if you are a small Python team shipping one agent. The five-language surface, the graph model, and the deployment tooling are all answers to enterprise problems, and you will pay for them in concepts you do not need — PydanticAI or the OpenAI Agents SDK will get you there faster. It is also the wrong pick if you are committed to another cloud and want the managed path: ADK will run on AWS or Azure, but Agent Engine is the part that makes the deployment story compelling, and that part is Google's.

Pricing and Licensing

ADK is free under the Apache 2.0 licence, with no paid tier of the framework, no seat charge, and no enterprise edition holding features back. Apache 2.0 is permissive enough for commercial use and redistribution, and it satisfies an open-source-only licence requirement — which is worth noting because it applies to the whole framework, including the graph workflows and the evaluation harness, not just a community core.

The costs sit in two places. The first is model spend, billed by whichever provider the agent calls. The examples use Gemini, and if you use Gemini, that is Google's per-token bill; ADK does not require it, and pointing an agent at another provider is a model string change.

The second is infrastructure, and it is the one to plan for. Running ADK yourself costs whatever your compute costs. Deploying to Vertex AI Agent Engine is a managed Google Cloud service with its own usage-based charges, separate from the model tokens it consumes. That is the honest comparison against a self-hosted framework: ADK's software is free at both ends, but the deployment path the product is designed around is a paid service.

Compared with the alternatives, the licensing is the most permissive position in the category alongside the OpenAI Agents SDK, PydanticAI, and LangGraph, all of which are free with no paid tier. CrewAI and LangChain are Freemium, with a paid platform beside the open framework. Budget for the model calls and, if you use it, for Agent Engine.

Strengths & Trade-offs

Pros

  • Five first-party language implementations — Python, TypeScript, Go, Java, Kotlin — which is rare and decisive for a JVM or Go shop.
  • Apache 2.0 across the whole framework, so no feature sits behind a plan and an open-source-only requirement is satisfied.
  • Graph workflows express a process that is partly deterministic, instead of forcing every step through a model.
  • A first-party evaluation harness, so "does this agent work" has an answer before deployment rather than after.
  • A managed deployment target in Vertex AI Agent Engine, which removes the runtime question for Google Cloud teams.

Cons

  • The centre of gravity is Google Cloud. The framework is portable; the deployment story that makes it attractive is not.
  • More concepts than a small team needs. Hierarchical composition, graph workflows, and the CLI are enterprise answers, and they cost learning time even when unused.
  • Version 2.x is recent, and the 2.0 graph-workflow redesign means older tutorials and examples describe a different framework.
  • The five language implementations do not all move together, so a feature announced for Python may not be in the Go or Kotlin build yet.
  • Documentation assumes Gemini in most examples, so using another provider means translating as you read.

Alternatives to Google Agent Development Kit

The reviewed substitutes for Google Agent Development Kit among the agent frameworks, and what would make each one the better answer.

Direct alternatives

Reviewed substitutes: products bought for the same job, where a team picks one.

LangGraph
Both express agent workflows as graphs and are adopted as the primary framework. ADK adds a managed deployment target; the framework decision is still one or the other.
OpenAI Agents SDK
The two large model vendors' own agent frameworks. A team building agents adopts one of them or a vendor-neutral framework; it does not adopt both.

Other approaches

A different approach to the same problem. Each substitutes only for the workload named beside it.

Semantic Kernel
Two platform-bound frameworks, and the condition is the cloud rather than the capability: Semantic Kernel is the Microsoft-stack answer for .NET and Azure shops, ADK the Google Cloud one, arriving with Vertex AI grounding and Agent Engine. A team evaluates whichever matches its estate and rarely both. Semantic Kernel is superseded: Microsoft merged it and AutoGen into the Microsoft Agent Framework, GA April 2026, and SK v1.x now receives critical bug and security fixes only, with new features built in the successor. It stays a real option because a .NET estate running it today is supported into 2027, and the catalogue records the supersession on its own page -- but the pair is a choice between a live framework and one on a known sunset, and saying so is the point.Applies to: Choosing an enterprise agent framework aligned to an existing cloud. Semantic Kernel stands in for ADK in a Microsoft estate already built on it; ADK stands in inside Google Cloud. A team starting fresh on Microsoft should be looking at the Agent Framework successor rather than either.
CrewAI
Both are on the same 2026 shortlist and win different shapes of work. CrewAI models a team of specialists with roles and tasks, which is chosen when the product team has to read the orchestration; ADK is chosen when the deployment target is Google Cloud and managed grounding and Agent Engine matter more than the authoring metaphor. Conditional because neither displaces the other outside its condition: CrewAI has no cloud-managed retrieval, ADK has no role-and-task abstraction.Applies to: Building a multi-step agent system. CrewAI stands in for ADK when the work divides into readable roles and the cloud is not fixed; ADK stands in when the team is on Google Cloud and wants managed grounding and deployment.
LangChain
Both sit on the 2026 shortlist a team draws when standardising one agent framework, and the guidance is explicit that there is no universal winner: the choice turns on workload shape, team language and governance. LangChain is picked for the broadest tool ecosystem and fast model swaps; ADK is picked when the estate is Google Cloud, because it arrives with Vertex AI grounding and Agent Engine deployment attached. Conditional rather than direct: ADK's advantage is the platform, so off GCP the two are not on the same shortlist at all.Applies to: Standardising one framework for agent development. ADK stands in for LangChain when the team is on Google Cloud and wants grounding and deployment in the box; LangChain stands in when the team needs provider portability or the widest tool ecosystem.
Agno
The fork is whether the agent runtime is tied to a cloud. Agno is model-agnostic and self-hosted, shipping its own memory and knowledge layer across 25+ vector stores and 29 embedders; ADK is Google Cloud's framework, arriving with Vertex AI grounding and Agent Engine deployment. A team that has not committed to GCP is choosing Agno for portability, and a team that has is choosing ADK because the managed pieces come free.Applies to: Building an agent with its own knowledge base. Agno stands in for ADK when the deployment must stay portable or self-hosted; ADK stands in when the team is on Google Cloud and wants managed grounding and deployment.
PydanticAI
Both are open-source Python agent frameworks with evaluation and deployment stories, and both sit in the same 2026 decision. The condition is the estate: ADK's case rests on Vertex AI grounding and Agent Engine, which are worth nothing off Google Cloud, while PydanticAI's case is idiomatic Python with typed tool interfaces and validated output, which travels anywhere. A team on GCP weighs the managed pieces; a team anywhere else is not comparing them at all.Applies to: Choosing a Python agent framework. ADK stands in for PydanticAI when the team is on Google Cloud and wants managed grounding and deployment; PydanticAI stands in when type safety and provider portability matter more than platform integration.
Haystack
Google's own framing is the choice: Vertex AI RAG Engine is a fully managed retrieval system handling upload, embedding, indexing and semantic search behind one API, against a pipeline you own and operate. Haystack is the own-it end of that -- every step serializable, testable and auditable with build-time validation -- and ADK is how a team builds the agent on top of the managed end. Conditional because the substitution holds only inside Google Cloud, and only where a managed pipeline is acceptable in place of an auditable one.Applies to: Grounding an agent in company documents. ADK with Vertex AI RAG Engine stands in for Haystack when the documents are standard types and the team wants turnkey retrieval on GCP; Haystack stands in when the pipeline must be inspectable and testable, or the team is not on Google Cloud.
LlamaIndex
Google's own docs frame this as the choice: Vertex AI RAG Engine is a fully managed retrieval system handling upload, embedding, indexing and semantic search behind one API, against a custom vector store built with LlamaIndex or LangChain where you own the pipeline. ADK is how a team builds the agent on top of that managed retrieval, with Agent Engine for deployment, so on GCP it covers the ingestion-to-agent span LlamaIndex covers everywhere. Conditional because the substitution holds only inside Google Cloud; off it, ADK is not on the shortlist.Applies to: Building and deploying an agent grounded in company documents on Google Cloud. ADK with Vertex AI RAG Engine stands in for LlamaIndex when the documents are standard types and the team wants turnkey retrieval without infrastructure work; LlamaIndex stands in when the team wants to own the pipeline or is not on GCP.
Explore all Google Agent Development Kit alternatives →

Public signals

About these signals

Verified factual signals from public sources. They indicate observable activity or interest, not total adoption, product quality, or cost.

1.3k GitHub commits 90d21.6k GitHub stars0 vulnerabilities across 1 package

See all signals from 4 sources
Source
Signals
Last updated
GitHub
Commits 90d:1.3kStars:21.6k
September 21, 2026
PyPI
Weekly downloads:2.4M
September 21, 2026
Product Hunt
Comments:5Rating:5.0/5Reviews:2Votes:140
September 21, 2026
OSV
Package vulnerabilities:0 vulnerabilitiesacross 1 package

PyPI · google-adk@2.9.2

September 21, 2026

Frequently asked questions

Is Google ADK free?

Yes. ADK is Apache 2.0 licensed with no paid tier and no feature held back for an enterprise edition. You pay for the models the agent calls, and for Vertex AI Agent Engine if you use that managed runtime rather than running ADK yourself.

Does ADK only work with Gemini?

No. Most documentation examples use Gemini, but the model is configuration and other providers work. The practical friction is that you translate examples as you read them.

Which languages does ADK support?

Python, TypeScript, Go, Java, and Kotlin, each with a first-party implementation. The implementations do not always move in step, so check that a specific feature has landed in the language you need.

ADK or the OpenAI Agents SDK?

Mostly a question of which platform you are on. Choose ADK for Google models, Gemini Enterprise Agent Platform (formerly Vertex AI) deployment, or a JVM, Go, or Kotlin codebase. Choose the OpenAI Agents SDK for OpenAI models, sandbox agents, and first-party realtime voice.

Related Agent Frameworks

Other agent frameworks in the catalog. Same kind of product, not a substitution recommendation.