Agent Run Cancellation

Cancel a running agent execution from another thread.

Cancel a running agent execution by starting the run in one thread and cancelling it from another. The example also shows how to handle cancelled responses.

Cancellation is cooperative: a request marks the run for cancellation, and execution stops at its next cancellation check. A fast run can finish first. The final status below determines the outcome; marking a run is not proof that it stopped.

Example

agent_cancel_run.py
"""
Example demonstrating how to cancel a running agent execution.

This example shows how to:
1. Start an agent run in a separate thread
2. Cancel the run from another thread
3. Handle the cancelled response
"""

import threading
import time
from uuid import uuid4

from agno.agent import Agent
from agno.models.openai import OpenAIResponses
from agno.run.agent import RunEvent
from agno.run.base import RunStatus


def long_running_task(agent: Agent, run_id_container: dict):
    """Consume the parent run's stream and retain its terminal outcome."""
    run_id = run_id_container["run_id"]
    status = "unknown"
    content_pieces = []
    try:
        for chunk in agent.run(
            "Write a detailed story about a dragon who learns to code.",
            run_id=run_id,
            stream=True,
            stream_events=True,
        ):
            # Member/step runs have their own IDs and do not determine parent status.
            if chunk.run_id != run_id:
                continue
            if chunk.event == RunEvent.run_cancelled:
                status = "cancelled"
            elif chunk.event == RunEvent.run_error:
                status = "error"
            elif chunk.event == RunEvent.run_paused:
                status = "paused"
            elif chunk.event == RunEvent.run_completed and status == "unknown":
                status = "completed"
            content = getattr(chunk, "content", None)
            if isinstance(content, str):
                content_pieces.append(content)
        run_id_container["result"] = {
            "status": status,
            "run_id": run_id,
            "cancelled": status == "cancelled",
            "content": "".join(content_pieces)[:200],
        }
    except Exception as exc:
        run_id_container["result"] = {
            "status": "error", "run_id": run_id, "cancelled": False,
            "error": str(exc), "content": "Run raised an exception",
        }


def cancel_after_delay(agent: Agent, run_id_container: dict, delay_seconds: int = 3):
    """
    Cancel the agent run after a specified delay.

    Args:
        agent: The agent whose run should be cancelled
        run_id_container: Dictionary containing the run_id to cancel
        delay_seconds: How long to wait before cancelling
    """
    print(f"Will cancel run in {delay_seconds} seconds...")
    time.sleep(delay_seconds)

    run_id = run_id_container.get("run_id")
    if run_id:
        print(f"Cancelling run: {run_id}")
        success = agent.cancel_run(run_id)
        if success:
            print(f"Run {run_id} marked for cancellation")
        else:
            print(
                f"Failed to cancel run {run_id} (may not exist or already completed)"
            )
    else:
        print("No run_id found to cancel")


def main():
    """Main function demonstrating agent run cancellation."""

    # Initialize the agent with a model
    agent = Agent(
        name="StorytellerAgent",
        model=OpenAIResponses(id="gpt-5.2"),  # Use a model that can generate long responses
        description="An agent that writes detailed stories",
    )

    print("Starting agent run cancellation example...")
    print("=" * 50)

    # Container to share run_id between threads
    run_id_container = {"run_id": str(uuid4())}

    # Start the agent run in a separate thread
    agent_thread = threading.Thread(
        target=lambda: long_running_task(agent, run_id_container), name="AgentRunThread"
    )

    # Start the cancellation thread
    cancel_thread = threading.Thread(
        target=cancel_after_delay,
        args=(agent, run_id_container, 8),  # Cancel after 8 seconds
        name="CancelThread",
    )

    # Start both threads
    print("Starting agent run thread...")
    agent_thread.start()

    print("Starting cancellation thread...")
    cancel_thread.start()

    # Wait for both threads to complete
    print("Waiting for threads to complete...")
    agent_thread.join()
    cancel_thread.join()

    # Print the results
    print("\n" + "=" * 50)
    print("RESULTS:")
    print("=" * 50)

    result = run_id_container.get("result")
    if result:
        print(f"Status: {result['status']}")
        print(f"Run ID: {result['run_id']}")
        print(f"Was Cancelled: {result['cancelled']}")

        if result.get("error"):
            print(f"Error: {result['error']}")
        else:
            print(f"Content Preview: {result['content']}")

        if result["cancelled"]:
            print("\nSUCCESS: Run was successfully cancelled!")
        elif result["status"] == "completed":
            print("Run completed before cancellation")
        else:
            print(f"Run ended with status: {result['status']}")
    else:
        print("No result obtained - check if cancellation happened during streaming")

    print("\nExample completed!")


if __name__ == "__main__":
    # Run the main example
    main()

Usage

Set up your virtual environment

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

Install dependencies

uv pip install -U agno openai

Export your OpenAI API key

Set OpenAI Key

Set your OPENAI_API_KEY as an environment variable. You can get one from OpenAI.

export OPENAI_API_KEY=sk-***

Run example

python agent_cancel_run.py

API Endpoint

Agent runs can be cancelled via the AgentOS API:

POST /agents/{agent_id}/runs/{run_id}/cancel

Start a separate AgentOS server with the entity registered before using this endpoint. The thread example above does not start an HTTP server. Replace the entity and run IDs below with IDs returned by that server and include its authentication headers when configured.

Example:

curl --location 'http://localhost:7777/agents/story-writer-agent/runs/123/cancel' \
  --request POST

Reference: Cancel Agent Run API