Carbide Create and the inaccuracy of selecting objects

Is there an official place to submit bug reports and software enhancement requests for carbide create?
I’m using it on a mac laptop and it is very frustrating. I am happy that the tool is free and if it wasn’t so tough to use I might consider paying for pro but doing simple things like selecting objects where there are several close objects is really frustrating. It’s easy to reproduce this problem. Put 5 boxes inside and near each other on the bed plane, then simply try to select the objects individually, often times the wrong box is selected. Now imaging trying to select 24 boxes to make a group and because its so inaccurate, you accidentally deselect everything. To make things worse, the undo hotkeys do not go back to the prior selection, and now you have just wasted a bunch of time trying to make a group. I hope this doesn’t come across as a rant. I just want to make sure that Carbide is aware of how hard it is to use the product on a mac laptop.

Which version are you using?

What size are the squares?

How zoomed in are you? What size is your working area?

I get an accurate highlighting for a row of five 20mm squares in my current project:

and was able to select each of the 5 at will.

Usually zooming in a bit will help.

Here’s an example of what I am working with.
Honesty the problem pops up with smaller jobs too. It feels like the pointer tool is just not accurate. It often just selects the wrong thing.

I’m working with a full 2ft x 4ft bed. Zooming in will work but its very hard to zoom in then out then in then out to try and select 24 boxes to group.

Screenshot 2023-10-22 at 11.21.35 PM

Jessie, let Control Z be your friend. I reckon I would be as mad as a cut snake if it wasn’t for that feature. Every time you hit the buttons, it goes back another step that you have made. It’s magic for an 80 year old such as me.
Have fun.
Cheers for now, Ben

This is more easily seen with nested objects. A huge part of the problem is that the ‘tip’ of the pointer is not the point of selection - instead, it seems that whatever object the pointer touches stays high-lighted for selection as long as any part of the pointer touches it, even if the pointer moves over other objects.

Of course I can’t take a screen shot of this behavior, since a screenshot removes the pointer :(.
In this image, the tip of the pointer was located outside the outermost box. It had started inside all the boxes, and I moved it towards the outside. But, since it was still touching something else from before, that is what got selected.

Yes. I understand the issue. I keep myself out of mischief by making a wooden clock. Wheels. pinions ,arms etc. Sometimes I have to zoom in a long way to be able to select a feature that I need, but zoom is very useful, especially for old eyes.
Cheers, Ben.

1 Like

I agree with @mhotchin that the selection is not at the tip of the pointer. As others suggested I use zoom a lot when working with small objects. For instance a .125" through hole surrounded by a 1/4" recess. It is very hard to select the 1/8" hole and keep getting the 1/4" recess. I never put 2 and 2 together but now that I think about it it is not the point of the curser that selects. I have always worked around it but it is frustrating.

This has gotten a bit better in the new beta:

If you find a particularly egregious instance of pointer position and selection/highlighting, draw in a circle to highlight where the pointer is (put it on a separate layer) and post that file and a screengrab here and we’ll see if we can get a developer to look into it.

Have you cut that file yet?

Are you using a laser, or do you square up the mortises with a file? Or do you relieve the tenons w/ a roundover tool?

Why no dogbones?

With a 1/8" bit on 1/8th luan ply, the tenans fit snugly without any filing. I think dogbone would take too much space for such small work.


The example I gave above with the nested squares was with the latest build (761).

Do you see the “pre-select” highlighting of the object? When you mouse over, the object that will be selected will get highlighted, then you can click.

I have found that I sometimes have to zoom in really tight to be able to differentiate the highlighting…I don’t try to rely on the pointer, only the highlighted object.


I do and honestly it’s slow. I move pretty fast and dealing with a lot of objects is not easy in the software. I really do wish they’d log it as bug. It would be great to see a github page where we can report real issues with the software.

Every issue reported here is considered and tracked, as is every issue e-mailed in to (or

How do we know if it’s accepted as a bug and how can we know if it’s coming in a future release? These types of decisions are useful as a user trying to decide which software to use when running a business.

1 Like

That’s all above my pay grade.

Our recommendation is that you evaluate our products based on what is currently available — we’re not fortunetellers and we don’t expect other people to be.

In addition to putting it here…try adding it as a Feature Request Here (I know, it’s a bug, not a feature…but an improved selection capability is a welcome feature).

I’ve found that Carbide addresses most defects quickly - particularly if there are no reasonable work-arounds. Nice-to-have’s have longer runways. I’m not sure where the powers that be will feel this one will fall. It’s probably one of those “#of people impacted vs. available resources vs. backlog of changes” questions.

  • G

Whoa! Was this at me? Sorry I’m not sure who you’re aiming at here.

1 Like

We use the drawing canvas provided by Qt, the cross platform framework we use. We’ve replaced a lot of their code by now, because its default behaviors are so bad for CAD applications. (Add it to the list of things we wish we knew from the start)

Selection is accurate to the tip of the pointer, as fas as I know, but when vectors are very close to one another then the selection system in the canvas might pick the first one drawn that is “close enough”, rather than the closest. We’re willing to dig in and add more code to change this, but we’d like to have some of your files to test with. If we make the wrong assumptions, this could add a lot of overhead.

Please post your files here, or send them to and we’ll dig into this more.

1 Like