Kiosk Mode in Windows – Chrome on Multiple Displays

The problem

Did you ever think about putting several TVs/monitors to the wall and show there some dashboards? Current status, some pie charts and trends, alerts or advertisements. Or much like an arrival/departure tables in airport?

So, how would you do it?

• “IT’s quite straightforward, isn’t it” – you would answer originally – “simply take a PC with necessary video slots, connect all monitors/TVs… and the thing’s done.”
• Really?
• “Ok, we would need some apps to show the information and most probably the is a portal will all that available, so simply launch several Chrome instances in full screen mode and that’s it“.

But if you are reading this blog post probably you’ve already realized that isn’t really so simple.

• First, because you want all instances to start in full screen mode automatically
• Second, you want them on certain monitors.

And when the first part is not really complicated – you may find chrome command line options allowing to start Chrome in full screen on even in kiosk mode – you will definitely face problems with the second one. Although Chrome command line has option for specifying window position, you will find out it doesn’t work in full screen, here is what you’ll find on the issue while googling:

• http://stackoverflow.com/questions/29294045/how-to-open-two-instances-of-chrome-kiosk-mode-in-different-displays-windows
• http://stackoverflow.com/questions/3750113/launch-an-application-and-send-it-to-second-monitor
• http://superuser.com/questions/879898/how-to-run-many-instances-of-google-chrome-in-full-screen-on-multi-monitor-syste?answertab=votes#tab-top
• http://superuser.com/questions/720917/starting-multiple-chrome-full-screen-instances-on-multiple-monitors-from-batch

But besides Chrome, you may want to launch some other app. Like video monitoring software (to ensure all slaves work hard J)

Enough Words, Give Me a Solution

A good start for the solution is given by the Stack Overflow answer.

As we need to be able to modify easily the apps, URL and entire solution easily, I’ve decided to create a script on PowerShell. But helper functions wrapping WinApi are put into a .NET dll.

A .NET/C# Library

Expose necessary WinApi functions:

public static class WinApi
{

    [DllImport("user32.dll", SetLastError = true)]
    public static extern bool SetWindowPos(IntPtr hWnd, IntPtr hWndInsertAfter,
           int X, int Y, int cx, int cy, SetWindowPosFlags uFlags);

    [DllImport("user32.dll")]
    [return: MarshalAs(UnmanagedType.Bool)]
    public static extern bool ShowWindow(IntPtr hWnd, ShowWindowCommands nCmdShow);

    [DllImport("user32.dll")]
    public static extern bool SetForegroundWindow(IntPtr hWnd);
}

I will not provide Enums referenced, you may find them on http://www.pinvoke.net/ on the respective functions description page, or download entirely in my code sample. (Link at the bottom)

Assuming we can start the process from a script easily, then next function we need is actually moving the window:

public static bool MoveToMonitor(IntPtr windowHandle, int monitor)
{
     monitor = monitor - 1;
     return WinApi.SetWindowPos(windowHandle, IntPtr.Zero, 
            Screen.AllScreens[monitor].WorkingArea.Left,
            Screen.AllScreens[monitor].WorkingArea.Top, 1000, 800, 
            SetWindowPosFlags.SWP_NOZORDER | SetWindowPosFlags.SWP_NOREDRAW);
}

But we cannot move full screen window easily this way. That’s why we will start Chrome in regular not full screen mode, and then send F11 (full Screen shortkey) to it via

public static void SendKey(IntPtr windowHandle, string keys)
{
         WinApi.SetForegroundWindow(windowHandle);
         SendKeys.SendWait(keys);
         SendKeys.Flush();
}

PowerShell Part

Before you start, ensure you have enabled scripts execution:

Run PowerShell as Administrator and execute the following, confirm when prompted.

Set-ExecutionPolicy RemoteSigned

Now let’s write a script implementing this logic:

First, let’s reference our assembly with helper functions described above. (These are class names from my real example, in your case names could be different.):

Add-Type -Path Tomin.Tools.KioskMode.dll
$WinAPI = [Tomin.Tools.KioskMode.WinApi]
$Helpers = [Tomin.Tools.KioskMode.Helper]

Second, here is the code

function Chrome-Kiosk($Url, $MonitorNum)
{

    Write-Output "starting chrome $Url , monitor: $MonitorNum"
    Start-Process $chromePath "$chromeArguments $Url"
    Start-Sleep -Seconds $ChromeStartDelay

    $window = (Get-Process -Name chrome | where MainWindowHandle -ne ([IntPtr]::Zero) | select -First 1).MainWindowHandle

    $WinAPI::ShowWindow($window, [Tomin.Tools.KioskMode.Enums.ShowWindowCommands]::Restore)
    $Helpers::MoveToMonitor($window, $MonitorNum)
    $Helpers::SendKey($window, '{F11}')
    Start-Sleep -Seconds $ChromeStartDelay
}

We moved it into a function to easy start several instances.

Let me explain what it does:

1. First we start a chrome instance with some command line parameters, an easy part.
2. Then we need a window handle to manipulate it. But Chrome is special, main window of the stated process is not the one we want. The good part – the is only one ManWindow set for all Chrome processes – that is one currently active. Thus we wait a second for chrome to create a window,
3. And look for that window handle, based on the above described criteria. Once window handle is know you are the master of it.
4. But as mentioned above there could be a problem trying to move full screened window, thus we restore it to normal
5. Move to required monitor (numbering starts from 1)
6. And send F11 key to it.
7. Last delay command is to wait for F11 to take effect, because otherwise next consequent function execution may fail, as Chrome did not switch his process to new window in time

Of course you will need also to initialize couple of variables:

$chromePath = 'C:\Program Files (x86)\Google\Chrome\Application\chrome.exe'
$chromeArguments = '--new-window --incognito'
# if Window not moved (especially on machine start) - try increaing the delay.
$ChromeStartDelay = 3

And the final calls would then look like:

# &taskkill /im chrome* /F
Chrome-Kiosk 'http://google.com' -MonitorNum 1
Chrome-Kiosk 'http://http://www.bbc.com/' -MonitorNum 2

The first command is optional and commented, but could be useful when somebody played with your kiosk and you want quickly to restore everything back by a single click.

Final Step

Create a link to your PowerShell script:

And put it into your Startup folder.

Entire Solution, No Programming

You may download already compiled solution with sample script.

Prerequisites

Windows 8.1

• No

Windows 7

• PowerShell 4: http://social.technet.microsoft.com/wiki/contents/articles/21016.how-to-install-windows-powershell-4-0.aspx
• .NET Framework 4.5: https://www.microsoft.com/en-US/download/details.aspx?id=42643

All you need to do:

1 Download and Extract

Download link is at the bottom.

Extract contents to C:\Kiosk-Mode.

You may put into another folder – but ensure you renamed path in all places.

2. Modify Ciklum-Kiosk.ps1

Put necessary URLs and Monitor Numbers (starting from 1)

3. Enable Scripts Execution

Run PowerShell as Administrator and execute the following, confirm when prompted.

Set-ExecutionPolicy RemoteSigned

4. Launch Shortcut

Double click on link (Start – Kios – modifypath.lnk) to ensure browsers are started and opened on necessary displays.

Note: youi may need to modify your link if you put file not into C:\Kiosk-Mode. Right click on it -> properties and fill it accordingly:

5. Autostart

Copy the link to desktop for easy access and to your startup folder:

6. Troubleshooting

If window is not moved to another desktop – try increasing the delay in script ($ChromeStartDelay) variable.

Script has been tested on Windows 8.1 machine.

Links

GitHub page: https://github.com/alex-tomin/Tomin.Tools.KioskMode

Binaries: https://github.com/alex-tomin/Tomin.Tools.KioskMode/releases/download/1.0/Tomin.Tools.KioskMode.zip