Augments ADK

A provider-agnostic Python framework to orchestrate complex systems of agents that perform real-world actions, across 100+ LLMs via LiteLLM.

• Foundations — The mathematical limits that shape every decision in this ADK.
• Architecture — Layer 1/2/3 types, Runner loop, LLM ABC, handoffs, swarms, graphs.
• Concepts — Every concept and how it differs from its neighbours.
• Guides — Practical how-tos: agents, tools, memory, guardrails, tracing.
• Topics — Deep per-module reference: graphs, swarms, tools, llms, sandbox, and more.
• CLI — Run, chat, validate, scaffold, and serve agents from the terminal.
• Deployment — Package, containerise, and ship agents to Kubernetes, Cloud Run, and AWS.
• References — API reference, contributing, changelog, maintenance.

Quickstart

uv add augments-adk    # or: pip install augments-adk
 
python -c "from augments.adk import Agent, Runner; print('OK')"

Working on the ADK itself instead? Clone the repo and run uv sync --extra dev (or conda env create -f environment.yaml) — see Contributing.

import asyncio
import logging
 
from augments.adk import Agent, Runner
 
logger = logging.getLogger(__name__)
 
agent = Agent(name="Assistant", system_prompt="You are a helpful assistant.")
result = asyncio.run(Runner.arun(agent, "Hello!"))
logger.info(result.final_output)

See Guides → Agents for the next step.