kneo-agent 1.0.1

Python SDK for building LLM-powered agents across multiple providers.

Kneo Agent

Kneo Agent is a Python SDK for building LLM-powered agents across multiple providers with one consistent API.

It supports:

• bridge-style runtimes where Kneo owns the execution loop
• native runtimes backed by provider SDK behavior
• adapter-style wrappers around existing provider objects
• workflows for orchestrating multi-step or multi-agent runs

Requirements

• Python 3.12+

Installation

Install the core package from PyPI:

pip install kneo-agent

Install with optional provider extras when needed:

pip install "kneo-agent[openai]"
pip install "kneo-agent[langchain]"
pip install "kneo-agent[google-adk]"
pip install "kneo-agent[all]"

Quick Start

import asyncio

from kneo_agent import AgentBuilder
from kneo_agent.patterns import BridgeAgentFactory


runtime = BridgeAgentFactory.for_openai(model="gpt-4o", strategy="react")

agent = (
    AgentBuilder()
    .with_name("Assistant")
    .with_system_prompt("You are a helpful assistant.")
    .use_bridge(runtime)
    .build()
)


async def main() -> None:
    reply = await agent.chat("What is 2 + 2?")
    print(reply)


asyncio.run(main())

Optional Features

• openai: OpenAI Agents integration
• langchain: LangChain integration
• google-adk: Google ADK integration
• all: all optional provider integrations

Package Contents

The published distribution contains the kneo_agent package and typed package metadata for downstream users.