By Grego
There’s a category of open source projects that appear on GitHub Trending and you immediately assume they’re a joke.
PeonPing falls squarely into that category.
Its pitch is ridiculous in the best possible way: add sounds from Warcraft III, StarCraft, Age of Empires, Portal and other classics to your coding agents.
Your agent finishes a task.
“Work complete.”
Finishes another.
“Job’s done.”
An error appears.
“You must construct additional pylons.”
The initial reaction is obvious: this is a meme.
The reaction after using it for a few days is different.
Because it accidentally solved a real problem.
The problem nobody was discussing
Coding agents changed the way we work.
Before, we wrote code, waited for compilations, ran tests and reviewed results manually.
Now we give a task to Claude Code, Codex, Cursor or OpenCode and go do something else.
The problem is that agents are silent.
Too silent.
You ask Claude Code to refactor twenty files. You open Slack while it works. You answer a message. You get distracted with another task. Ten minutes later you realize the agent finished nine minutes ago and has been waiting for instructions the whole time.
If this happened to you, you’re not alone.
It’s one of the small friction points that appear when agents start becoming part of your daily workflow.
It’s not a technical problem.
It’s an attention problem.
What PeonPing does
PeonPing is a minimal layer of sound notifications for development agents.
It listens for events from tools like:
- Claude Code
- Codex CLI
- Cursor
- OpenCode
- Kiro
- Antigravity
And plays sounds when important actions occur:
- task completed
- error
- approval requested
- intervention required
- execution finished
The trick is that the sounds aren’t generic.
They’re direct references to decades of strategy games and PC gaming.
The result is strangely effective.
Your brain instantly identifies that something happened without forcing you to look at a window, a terminal, or an operating system notification.
The fun part
Some of the available notifications include:
- Warcraft III
- StarCraft
- Age of Empires
- Portal
- Half-Life
- Command & Conquer
And yes, there are people setting up different agents with different voices to know which one finished first.
It’s not strictly necessary.
But neither was putting RGB lights on a keyboard and look where we are.
The surprising part: it actually improves workflow
After the initial novelty, something interesting happens.
Agents are asynchronous processes.
And humans are bad at monitoring asynchronous processes.
Every time you have to manually check if an agent finished, you’re paying a small cognitive penalty.
It seems insignificant.
But multiply it by twenty or thirty executions per day.
Suddenly you understand why a well-designed audio notification can feel more useful than it should.
It’s the same reason IDEs have sounds, compilations have alerts, and operating systems have notifications.
The difference is that until now, coding agents didn’t have a good solution for this.
PeonPing fills that gap with a surprisingly small amount of code.
Installation
Installation is straightforward.
With Node.js installed:
npx peonping
Or globally:
npm install -g peonping
Once started, it detects and monitors compatible agents in the background.
The configuration allows you to select different sound packs and assign sounds to specific events.
For example:
peonping config
From there you can choose:
- Warcraft mode
- StarCraft mode
- Age of Empires mode
- Mixed mode
- Custom sounds
Is it worth it?
If you run one or two prompts a day, probably not.
If you work eight hours alongside Claude Code, Codex or Cursor, surprisingly yes.
PeonPing looks like a joke because it was born as a joke.
But many of the best productivity tools are born exactly that way: someone spots a small annoyance, builds an absurd solution, and ends up discovering the problem was more real than it seemed.
It won’t write better code.
It won’t improve your prompts.
It won’t replace any AI tool.
But the next time you hear a “Job’s done” while you’re checking Slack and realize your agent finished three seconds ago instead of ten minutes ago, you’ll probably understand why this repo ended up on GitHub Trending.
And why the most absurd tool of the week might be one of the most useful.