Unattended robot on RDP. Can it use full UI recognition, or only image?



I’m attempting to use unattended robots without Orchestrator.

  1. How do I start or queue processes or even add them so they are available? I found this

If your Robot is not connected to Orchestrator, the available processes displayed are the ones that are saved in the %ProgramData%\UiPath\Packages and %ProgramData%\UiPath\Projects folders.

I put my projects in there, but they aren’t showing up.

  1. Can the desktop processes I’ve created (uses UI recognition) work on an unattended robot on a remote desktop, or does the image recognition only work?



I can give a little insight on number 2…
(Assuming the process is Run directly on the remote machine) Processes that require element positions on the screen and image size will only work with the RDP window open and logged in with the same resolution that you designed the process with. If you minimize or close the RDP while the process is running it will fault, however, works fine if you just keep the window in the background without minimizing. Most UI recognition that doesn’t involve size and location will work fine.

Now, if you want the size and positions to work correctly with a Back-office unattended robot, then you can use the Invoke Interactive which let’s you choose the Color depth and resolution, but kicks you off the machine when it starts (which you can log back into, provided you do so with correct resolution, to attend the robot)

Hopefully that makes sense, from my experience.