Thorsten Ball//August 25, 2026
Here are some of the letters we received in the last weeks:
“Orb? What’s an orb?” “Ralph, loops, graphs, and now orbs?” “What the fuck are orbs?” “You can’t call them orbs” “I hate it when people say ‘it’s orbin’ time’ - I hate it when people have fun!”
Okay, that last one’s made-up, threw it in there as a joke.
You gotta keep your humour when people get offended by names.
And at the same time, it’s easy to laugh because we also get many letters like this one:
But how do I make you go from angry to happy? I’d say: try orbs! Go to ampcode.com and try them! But it looks like the name’s the problem, so, let’s talk about names.
Amp is the name of our company and our product: a coding agent and a development environment.
Orbs are part of our product: they are remote agents — coding agents that run in a remote machine (i.e. not your local computer) — and you can remote control them from ampcode.com, from your phone, or from our CLI.
Yes, they are VMs, or sandboxes, or whatever you want to call them. Remote machines that we spin up for you, put your code into, and start the Amp agent on. They come in different sizes.
But with orbs you don’t really have to think about the “remote” or the “machine” parts.
Here’s how you run an agent in an orb:
New thread
and type in your promptAn orb will spawn up, with your code on it, and the agent will start running.
This is what it looks like, uncut:
Once the agent is idle, the orb will go to sleep. When you send another prompt, it will wake up again and so too will the agent. (Orbs can sleep for as long as you want: we’ve had orbs asleep for weeks and then came back to them. Doesn’t cost anything. You only pay when the orb is running.)
You can watch and control the orb from anywhere: on your phone, on the web, in the TUI.
You can spawn hundreds of them. Thousands. There’s no limit.
You can choose between small and large orbs.
That’s it, now you know what orbs are.
Except…
When people refer to orbs and say that they’re “magical”, or that they completely left behind their local development setup, or even say stuff like “it’s orbin’ time!” what they mean is more than that. Orbs are more than just VMs.
Orbs have portals, for example. Anything that can speak HTTP and listen on a port can be a portal to your orb, and you can open it in Amp:
You can annotate the things you see in a portal and send messages to the agent:
Each portal has a custom URL that you can share with your teammates in case you want their feedback on what they’re working on.
Speaking of URLs: the whole orb thread (i.e. your conversation with the agent in the orb) has a URL and you can share that too. Your teammates can then see the whole conversation you had with the agent:
And then, guess what, you can also invite your teammates to the orb and have multiplayer orbs in which you can chat with your teammates and they can chat with your agent:
Of course you can review what the agent built in that orb:
You can browse through other files in your repository:
You can use a terminal in case you want to execute something on the orb:
And, remember when I said that orbs go to sleep when the agent is idle? You can wake it on a schedule. Every 30 minutes, every morning at 9am, or every Sunday, whatever:
Just say something like “Do this every 45 min for the next 8 hours” and the agent will obey and will set a schedule and the orb will go to sleep and wake up again in 45 min.
But not only can we ask an agent to wake itself again, no. We can also ask the agent to spawn other orbs:
And agents can send messages and files to each other, inside orbs:
Combine all of this and use it for a few minutes and you realize: oh yeah, orbs are more than just VMs, it’s a whole thing.
And that’s why we called them orbs and not just “remote agents” or “cloud agents” or “agent that runs in an ephemeral cloud box that wakes up and goes to sleep depending on the agent’s state and that has a unique URL and some very sweet features around it.”