Every change clicked through and tested

Your project’s setup installs dependencies when a machine is created, and its startup entries start your databases and dev servers on every boot. An entry with a port is published as a preview you can open from the thread. You can also ask Capy to write the setup for you.

Yes. Every cloud machine has a desktop with Google Chrome, streamed to the thread pane, so Capy clicks through your app while you watch. It can attach screenshots or a video of the run to the pull request.

Yes, on Capy Pro, Max and Enterprise. Capy builds your app on Revyl’s build runners, installs it on a cloud iPhone, iPad or Pixel 7, and taps, swipes and types through it while it reads the app’s logs. For the iOS Simulator, run the thread on your Mac with the desktop app.

Open the machine’s desktop or a phone’s tab from the thread pane. In a phone’s live view you can tap, swipe and type yourself, and you can tell Capy in the chat to wait while you use it.

A background command runs for 30 minutes by default and up to 6 hours, and the machine stays awake while Capy waits on it. For longer jobs, Capy splits the work into shorter commands and saves results between them.

Yes. Capy splits the work across parallel agents, each on its own machine, so every variant runs on its own VM. Machines go up to 16 vCPU and 64 GB on Max and Enterprise, and your plan sets how much memory can run at once.

Yes. An automation runs a stored prompt on a cron schedule, on a GitHub or Slack event, or when your deploy posts to its webhook URL. Every run gets a full development environment, so Capy can run the tests and open a pull request.

Machine time bills Capy credits only while a machine is awake, from $0.10 an hour for Small to $1.60 for Hyper. A cloud iPhone or iPad bills $5.40 an hour and an Android phone $3.60.