Cometapi Basic

Run a GPT model through the CometAPI gateway in sync, async, and streaming modes.

basic.py
"""
Cometapi Basic
==============

Cookbook example for `cometapi/basic.py`.
"""

from agno.agent import Agent, RunOutput  # noqa
from agno.models.cometapi import CometAPI
import asyncio

# ---------------------------------------------------------------------------
# Create Agent
# ---------------------------------------------------------------------------

agent = Agent(model=CometAPI(id="gpt-5.2"), markdown=True)

# Get the response in a variable
# run: RunOutput = agent.run("Explain quantum computing in simple terms")
# print(run.content)

# Print the response in the terminal

# ---------------------------------------------------------------------------
# Run Agent
# ---------------------------------------------------------------------------
if __name__ == "__main__":
    # --- Sync ---
    agent.print_response("Explain quantum computing in simple terms")

    # --- Sync + Streaming ---
    agent.print_response("Explain quantum computing in simple terms", stream=True)

    # --- Async ---
    asyncio.run(agent.aprint_response("Share a 2 sentence horror story"))

    # --- Async + Streaming ---
    asyncio.run(
        agent.aprint_response(
            "Write a short poem about artificial intelligence", stream=True
        )
    )

Run the Example

Set up your virtual environment

uv venv --python 3.12
source .venv/bin/activate

Install dependencies

uv pip install -U agno openai

Export environment variables

export COMETAPI_KEY="your_cometapi_key_here"

Run the example

Save the code above as basic.py, then run:

python basic.py

Full source: cookbook/90_models/cometapi/basic.py