# Stop babysitting your AI agents in the terminal — NockIt sends push alerts straight to your phone (Nockit)

> Source: <https://dev.to/codecrack01/stop-babysitting-your-ai-agents-in-the-terminal-nockit-sends-push-alerts-straight-to-your-phone-3406>
> Published: 2026-08-05 08:14:17+00:00

If you're running autonomous AI agents, fine-tuning scripts, or long background jobs, you know the drill: you hit `enter`

, switch to another tab to work on something else, and then spend the next hour periodically flipping back to the terminal just to check if:

Watching lines of terminal output scroll by isn't a good use of anyone's time.

To solve this for my own workflow, I built **NockIt**—a free, open-source, plug-and-play notification utility that hooks into your local or remote AI agents and pushes real-time alerts directly to your phone or desktop via ** ntfy**.

**Learn More:** [https://nockit.uk](https://www.google.com/search?q=https://nockit.uk)

NockIt provides a simple, zero-bloat bridge between your autonomous dev pipelines and your devices.

`ntfy`

:Instead of keeping your terminal pinned to your screen while your agent runs through an multi-step workflow, you can trigger a push alert when key state changes happen:

``` python
import requests

def notify_dev(title, message, priority="default"):
    requests.post(
        "https://ntfy.sh/your_custom_nockit_topic",
        data=message.encode('utf-8'),
        headers={
            "Title": title,
            "Priority": priority,
            "Tags": "robot,warning"
        }
    )

# Inside your agent's execution loop:
try:
    agent.run_task()
    notify_dev("Agent Finished!", "Task completed successfully without errors.")
except Exception as e:
    notify_dev("Agent Crashed 🚨", f"Error encountered: {str(e)}", priority="high")
```

Ofc, no one wants to write such long code! Just set it up via a single command (requires Node Js installed):

```
npx -y nockit-mcp@latest setup
```

I'd love for the community to test it out and share feedback!

Drop a comment below and share your experience—what's your current setup for monitoring long-running background tasks or AI scripts, and what features would make your workflow easier?