The why / A note from Malkio

Over and out is not the end of the conversation.

I built chrrpr for me first.

I'm an AI power user with actual ADHD. I run dozens of agents, across harnesses and threads. They're good at working. They're bad at reaching me. A wall of text in a terminal I'm not looking at isn't a conversation. It's a missed call.

My kind of neurodivergence works far better with sound. A notification is easy to ignore. I don't need to be notified. I need to be engaged: something that turns my head and brings my attention back to where it needs to go. That's what gets me to follow through.

So I gave my agents a voice.

When an agent needs a decision, it hails me: short, by name, radio-style. I answer by holding a button and talking. My answer goes back into the exact session that asked, and the work keeps moving.

The convenience comes second, but it's real. At the desk, I stop babysitting tabs. Away from it, I can still unblock a build, approve a step, or say "not now" from a walk, a drive, the gym. Touch grass, stay on net.

I use chrrpr every day, in more ways than I planned. I think you might too.

chrrpr isn't another agent. It's the counterpart to every autonomous agent: the channel that lets them speak up when they need you, and lets you stay in charge without staying in the chair.

It's for people who:

• live where heavy AI use meets real ADHD, or any brain that follows through better when it's spoken to
• run more than one agent, or more than one harness, at the same time
• are tired of checking terminals to find out they were needed an hour ago
• want their agents to ask less often, but at the right moment
• want to step away without the work stopping

I'm building this for me, and I'll keep building it either way. If enough of you want it, I'll put more into the parts that make it work for others, shaped by your input. If not, that's fine too. The channel stays open.

— Malkio