A code-first Python framework for building, evaluating, and deploying AI agents using a graph-based execution engine. It supports structured delegation and deterministic execution flows but requires external generative AI models for inference.
Project overview
The framework enables developers to compose deterministic execution flows for agentic apps with support for routing, fan-out/fan-in, loops, and state management.
Project type
AI Agent · Workflow · Infrastructure
Use cases
Coding & Development · Automation
Deployment
Refer to project documentation
License
Apache-2.0
Best for
Developers and AI engineers who need a code-first Python framework to build, evaluate, and deploy sophisticated AI agents with structured delegation and deterministic execution flows.
Teams building AI agent automation prototypes that require a graph-based execution engine supporting routing, fan-out/fan-in, loops, and state management.
Key capabilities
A graph-based execution engine for composing deterministic execution flows for agentic apps, supporting routing, fan-out/fan-in, loops, retry, state management, dynamic nodes, human-in-the-loop, and nested workflows.
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.
Class for defining an AI's instructions, tools, and behavior.
Run agents locally using a CLI interface.
Run and test multi-agent directories or single agent folders using a local Web UI.
Limitations and risks
ADK 2.0 introduces breaking changes to the agent API, event model, and session schema, resulting in incompatibility with older 1.x versions.
Getting started
Install the framework easily via the standard pip command (pip install google-adk).
After installation, use the provided direct CLI or local Web UI commands to run and test single agent folders or multi-agent directories locally.
Alternatives and comparisons
An open-source agent harness that researches, codes, and creates using sub-agents, memory, and sandboxes to perform complex long-horizon tasks.
An open-source operations platform enabling engineers to build AI Agents, Apps, and Automations that take action across business systems via a visual builder.
GitHub project description: An open-source, code-first Python toolkit for building, evaluating, and deploying sophisticated AI agents with flexibility and control.
README: An open-source, code-first Python framework for building, evaluating, and deploying sophisticated AI agents with flexibility and control.
README: Workflow Runtime: A graph-based execution engine for composing deterministic execution flows for agentic apps, with support for routing, fan-out/fan-in, loops, retry, state manage…
README: 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…
README: > **Beginner Note:** ADK applications are built using two main classes: > **`Agent`** (defines an AI's instructions, tools, and behavior) and > **`Workflow`** (orchestrates agents…