Streaming-first runtime
Run and stream through one execution pipeline instead of maintaining separate runtime paths.
Open-source Python AI agent framework
Agiwo is a streaming-first Python AI agent SDK and control plane for teams building production-grade agents with tools, scheduler orchestration, tracing, and persistence.
Run and stream through one execution pipeline instead of maintaining separate runtime paths.
Combine builtin tools, custom tools, and agent-as-tool composition behind a stable contract.
Coordinate roots, child agents, waiting, wakeups, steering, and lifecycle through one owner.
Persist runs and steps, query traces, and inspect execution after the fact.
Quick example
The public site should immediately show that Agiwo is practical, not just conceptual.
import asyncio
from agiwo.agent import Agent, AgentConfig
from agiwo.llm import OpenAIModel
async def main() -> None:
agent = Agent(
AgentConfig(
name="assistant",
description="A helpful assistant",
system_prompt="You are a concise assistant.",
),
model=OpenAIModel(id="gpt-4o-mini"),
)
result = await agent.run("What is 2 + 2?")
print(result.response)
await agent.close()
asyncio.run(main())
Architecture
Agiwo keeps the runtime boundaries explicit so execution, orchestration, and observability stay understandable as systems grow.
Owns prompts, execution, hooks, runtime context, and step lifecycle.
Defines stable interfaces for builtin, custom, and nested-agent tools.
Coordinates root and child execution, waiting, routing, and wake conditions.
Projects runtime state and traces without becoming execution truth.
Comparison
Use a problem-oriented comparison instead of evaluating each project in isolation. The first comparison page focuses on orchestration, runtime boundaries, and observability.
Read the comparisonFAQ
Agiwo is an open-source Python AI agent framework and control plane focused on streaming execution, tool use, orchestration, tracing, and persistence.
It is for developers building production-grade LLM agents and multi-agent systems in Python.
Yes. Agiwo supports both agent-as-tool composition and scheduler-driven orchestration for longer-lived workflows.