you will find the Mac OS versions here:
https://github.com/ohglobal/NDI-Classify/releases
scroll down to find 1.1 Update (Mac OS) and then click on Assets

Does anyone have any experience using the tool NDI-Classify by Andy Carluccio? He did a tutorial on Troikatronix website about using trained ai models to send image recognition data to Isadora via a program he made called NDI-Classify.
*Tutorial can be found here https://support.troikatronix.com/support/solutions/articles/13000091703-teaching-isadora-to-recognize-images-objects-and-people
The program is for both windows and mac but the tutorial only walks through using NDI-Classify in the context of windows, opening an .exe file in the process where on my mac there is no .exe file. I see a .py file, so i have been trying to launch it through python launcher or IDLE with no luck. If anybody has any advice on this subject I would greatly appreciate it!
The Enttec DMX control actor is a little difficult to set up, but if you have the correct channels set, then it should function in the same way as a lighting desk. The times when I had a problem controlling lighting through the Enttec was fixed by experimenting with different speeds in the Serial Port setup - sometimes the default speed is too fast for the device.
I'd recommend building your file using the OSC Address Listener actor.
That way, whatever tool you made to do this can be put into a User Actor and used in any file instead of being locked to this one specific Isadora file because of Stream Setup.
@mark_m said:
Thoughts? Explanations? Logic?
It's showing the size of a blank character (space). Notice that if you leave the 'text' field blank or put a space in it, the text width/height doesn't change. Seems like a bug. I'll log it.
@mark_m said:
I can get around this using a limit-scale-value actor, but still curious as to why this is happening...
Here's another (admittedly more complicated) way (check the top right Comment): dynamic-subtitle-background-2026-04-24-4.1.3.izz
I am using the draw text actor's measure mode to scale a shape which is the background of the text. Like the background for subtitles.
For some reason, even when there's no text input, the text width and text height have values, when I would expect them to be zero.
I
I can get around this using a limit-scale-value actor, but still curious as to why this is happening...
Thoughts? Explanations? Logic?
Mark (not that Mark, who would know why!)
Thank you! I was able to use the stream settings to help troubleshoot my issue. The data was coming in under channel 0 in the stream settings viewer, once I changed that it is running fine.
By default Isadora uses port 1234 for OSC. This can be changed in the Isadora Preferences if you can't set it in the app sending the data.

You can use the stream setup dialog to detect incoming values. (located under Communications/Stream Setup). By checking Auto-detect Isadora will list all incoming values. This makes it easier to tell if you app is sending in an acceptable way, or needs some config changes.

Any incoming values will be listed in the blank space, with a few settings. The Channel number will be used to grab this incoming value in the Isadora patch.

Here the OSC Listener is used to 'grab' the OSC value. Channel '1' is set to grab the first address defined in the 'Stream Setup' dialog.
Hello,
I am trying to send position data from a scenic automation software (Creative Connors Spikemark) via OSC to Isadora. My goal is to have a projected image follow the automated scenery. I was not able to find a plug in or OSC dictionary for Isadora to receive the position data. Is there a way anyone has found to do this?
Thanks!
-J
Sales & Licensing questions are better to ask using the support ticket link in my signature.
Our S&L staff don't use the forum.