Blue Iris Windows 11 or Windows Server?

[sprucecams]

My installed version of BI 5 was at most 2 updates old so I cannot believe that was an issue. In any event, I sent that same report to BI Support and received a reply from "Jennifer" who stated the following:

Sometimes this happens if you previously had a very old version 5 on the system.
Run REGEDIT and find the key

HKEY_LOCAL_MACHINE\SOFTWARE\Perspective Software\Blue Iris

If you see a subkey 'a' beneath this, it must be deleted. You may need to right-click and change permissions on the key to allow "everyone" full control of the key before it may be deleted.

I did as she advised and my BI5 install booted up just fine with all my camera settings seemingly imported accurately. :-D I will be monitoring this for a few days before moving the new PC into my wire closet. So far so good!

 

[Flintstone61]

Just a simplemans wild guess
Thankd for sharing the solution....I have 1 machine here with early ver 5 and Win 10...if I were to try that trick..( importing a .reg file) ..from W11 back to 10with diff versions...i could see the same result...

 

[Pentagano]

Last night I moved my BI on dedicated pc with Win10 to a VM on Unraid server using Windows 11 tiny iso.
After some initial issues with getting the GPU passthrough seems to be going ok.

But I cannot seem to get yolo 6.2 cuda working with the GPU. Worked fine on the win10 machine.
So on the 5.2 net at the moment. My gtx970 worked well with the 6.2.
But still 40-50ms inference speeds on large model with high res jpegs is not too bad for the old gtx970.

 

[Pentagano]

10 cameras running windows 10 enterprise SPECTRE light. Very stripped down version of windows - VM has 2GB of ram assigned but only uses 1.2GB RAM!!!
Rock solid

 

[PabloTodd]

It sounds like you’ve been making some solid progress with your setup! I had a similar issue when I moved my setup to a VM and had trouble with CUDA not working right. I ended up doing a factory reset Windows to get a clean slate, which fixed a bunch of glitches and got things working smoothly again. Sometimes, starting fresh can help clear up those unexpected problems. If you’re still having issues with CUDA and GPU passthrough, it might be worth checking if there are any updates or specific settings for the VM. Your GTX970 still seems to be holding up well with those speeds!

 

[Jagradang]

I installed BI on windows server 2022 - and it seems to be fantatic. Everything was great until...I got to docker. Apparently docker desktop is not suoported on windows server. I run CodeProjects AI in docker as its much easier to manage and upgrade as required. I can get it installed but it won't start on boot therefore i have to manually log in everytime a reboot/crash or anything happens... Or i won't get any alerts! which is not good!. Currently trying to find any/all work arounds for this. If i find anything i'll write up on here, if not then i may have to switch to bloatware version of windows (win 11 pro) which is a shame!!

 

[iwanttosee]

I installed BI on windows server 2022 - and it seems to be fantatic. Everything was great until...I got to docker. Apparently docker desktop is not suoported on windows server. I run CodeProjects AI in docker as its much easier to manage and upgrade as required. I can get it installed but it won't start on boot therefore i have to manually log in everytime a reboot/crash or anything happens... Or i won't get any alerts! which is not good!. Currently trying to find any/all work arounds for this. If i find anything i'll write up on here, if not then i may have to switch to bloatware version of windows (win 11 pro) which is a shame!!

I use Proxmox to run Windows VM to run BI and Debian VM to host Docker. Works great.

 

[BobbyArcher]

My BI computer is running Windows 10 IoT Enterprise LTSC 21H2 since 04/2023 without any issues.

 

[TobiasF]

I run CodeProjects AI in docker as its much easier to manage and upgrade as required.

Interesting news. For testing purpose I was running CodeProjects AI separatly under Proxmox in a Linux LXC with Docker in it.
My plan for the target platform was running both Blue Iris and CodeProjects AI on the same windows machine to make use of the GPU by both programs.

I use Proxmox to run Windows VM to run BI and Debian VM to host Docker. Works great.

Do you use a separate GPU for Blue Iris and CodeProjects AI? Or only one gets it?
If Blue Iris would exist for Linux there would be much better options for virtualization with shared hardware.

I installed BI on windows server 2022 - and it seems to be fantatic.

It also works great under Server 2025. I could not notice a difference when testing on Windows 11 24H2 and Server 2025.
From technical side, the server Windows I see as the better platform for such kind programs. Less "bloatware". From license point of view a different topic.