Using in Jupyter Notebooks

PyFly Agentic works seamlessly in Jupyter and JupyterLab. Because the framework is async-first, use await directly in cells — Jupyter provides a running event loop automatically.

Setup

cd fireflyframework-pyfly
source .venv/bin/activate            # the venv with .[agentic] installed
python -m pip install ipykernel jupyterlab
python -m ipykernel install --user --name pyflyagentic --display-name "PyFly Agentic"
jupyter lab                          # or: jupyter notebook

Then select the PyFly Agentic kernel when creating a new notebook.

Example notebook

# Cell 1 — configure
import os
from getpass import getpass
from dotenv import load_dotenv

load_dotenv()
if not os.environ.get("OPENAI_API_KEY"):
    os.environ["OPENAI_API_KEY"] = getpass("OpenAI API key: ")
os.environ["PYFLY_AGENTIC_DEFAULT_MODEL"] = "openai-responses:gpt-6-luna"

# Cell 2 — create an agent
from pyfly.agentic.agents import Agent

from pyfly.agentic.models import ModelOptions

agent = Agent(name="notebook-bot", model_options=ModelOptions(max_tokens=4096))
result = await agent.run("Explain quantum entanglement in two sentences.")
print(result.output)

# Cell 3 — memory for multi-turn conversations
from pyfly.agentic.memory import MemoryManager

memory = MemoryManager(max_conversation_tokens=32_000)
agent_with_mem = Agent(name="chat", memory=memory)

cid = memory.new_conversation()
await agent_with_mem.run("My name is Alice.", conversation_id=cid)
result = await agent_with_mem.run("What is my name?", conversation_id=cid)
print(result.output)  # Alice

# Cell 4 — reasoning patterns
from pyfly.agentic.reasoning import ReActPattern

react = ReActPattern(max_steps=5)
result = await react.execute(agent, "What are the top 3 uses of Python in 2026?")
print(result.output)

# Cell 5 — structured output with validation
from pydantic import BaseModel
from pyfly.agentic.validation import OutputReviewer

class Summary(BaseModel):
    title: str
    bullet_points: list[str]
    confidence: float

summary_agent = Agent(name="summary", output_type=Summary)
reviewer = OutputReviewer(output_type=Summary, max_retries=2)
result = await reviewer.review(summary_agent, "Summarize the benefits of async Python.")
result.output  # displays the structured Summary object in the notebook

Tip

You do not need asyncio.run() or nest_asyncio in Jupyter — await works at the top level of any cell because Jupyter runs its own event loop.

Notebook runtime ownership

Create one AgenticRuntime(config=AgenticConfig()) per notebook session, activate it around agent construction and work, and await runtime.aclose() when finished. Re-running creation cells can register a name twice; close the prior runtime or use auto_register=False. See native lifecycle.