Run enterprise grade agent systems in your cloud with the AgentOS. Join the waitlist.
import json
from textwrap import dedent
import httpx
from agno.agent import Agent
from agno.models.openai import OpenAIChat
def get_upcoming_spacex_launches(num_launches: int = 5) -> str:
url = "https://api.spacexdata.com/v5/launches/upcoming"
launches = httpx.get(url).json()
launches = sorted(launches, key=lambda x: x["date_unix"])[:num_launches]
return json.dumps(launches, indent=4)
# Create an Agent that has access to real-time SpaceX data
agent = Agent(
model=OpenAIChat(id="gpt-4.1"),
# Each function in the context is evaluated at runtime
context={"upcoming_spacex_launches": get_upcoming_spacex_launches},
description=dedent("""\
You are a cosmic analyst and spaceflight enthusiast. 🚀
Here are the next SpaceX launches:
{upcoming_spacex_launches}\
"""),
# add_state_in_messages will make the `upcoming_spacex_launches` variable
# available in the description and instructions
add_state_in_messages=True,
markdown=True,
)
agent.print_response(
"Tell me about the upcoming SpaceX missions.",
stream=True,
)
Create a virtual environment
Terminal
and create a python virtual environment.python3 -m venv .venv
source .venv/bin/activate
Install libraries
pip install -U agno httpx
Run Example
python cookbook/agent_concepts/context/03-context_in_instructions.py
Was this page helpful?