Zach Lloyd

With warps, like standard agent, the standard mode of execution is it happens locally on your laptop.

Yeah.

Basically, yes.

So it's super flexible.

So you can configure Oz agents to run as part of your CI flows, like you said.

So if you want an agent that runs when you push code, you can have an agent that does code review.

You can have an agent that does security scanning, anything like that.

But you can also run these agents ad hoc.

So actually directly from Warp.

For your example, if you wanted to...

You don't have to run that agent locally, that first one that you talked about.

You could actually just fire it off in the cloud.

And the cool thing about running it in the cloud is it runs even if you close your laptop, you can check it from your mobile phone.

You can run an unlimited number of them without it clogging up your machine, which is a problem that we're seeing people have more and more where it's like people are multitasking so much with agents that they're running out of disk space and RAM and stuff.

And then the other thing you can do with these is like, you don't have to put them in GitHub or Slack or like you can, but they're fully programmable.

So if you like literally wanted to set up a system where it's like anytime, you know, something happens in your web app, you run an agent to do something.

There's like a REST API and an SDK for doing that.

So it's a super flexible way of taking like the coding agent that you have in Warp

And just being like, okay, I want to like program against it or have it automate stuff.

That's the general concept.