You can't do things like speech rec using a microcontroller.... You need the power of a PC.... So to answer your question the ezb4 needs to be connected (via wifi) to your PC at all times.... Code is "not" download to the ezb4 like it is with the Arduino... Code is executed on your PC and instructions are sent and received over wifi in real time to the ezb in order execute your code.... This happens so fast you wouldn't know that no code was ever downloaded to the ezb... The PC does all the work.... and remember all these powerful features like voice rec, camera tracking, object recognition, etc can't be done without a PC.... Because of this relationship between the ezb4 and computer the ezb4 robot controller is 10 to 100 times more powerful than an Arduino....

Thank you for your response, Richard. May I follow up with 2 more questions:

1. So wireless tethering only works with a PC/MAC? Or will it work with an Android mobile device?

2. For vision recognition, is the camera hooked up the robot kit? Or the host PC? If it is hooked up the robot, is the video information from the camera wirelessly transmitted back to the host PC for processing?

Unfortunately not Mac, but mobile devices like Android and IOS will work.... But be aware.... As far as I know voice rec is still PC only... A lot of us including myself use small windows 8" tablets like the acer w3 which works excellent with ARC...

You can use the ezb4 camera (plugs into the ezb4 for super convenience) or you can use any camera of your choice as long as it is recognized by Windows... You would have to power it separately as only the ezb4 camera plugs directly into the ezb4... FYI the ezb4 camera is pretty good and priced reasonably...

For iOS and Android, you develop the robot program on the PC, upload it to the EZ-B cloud and download to your device to run on the mobile version. As Richard said, voice recognition and video processing are not yet supported on the Mobile version.

ARC also runs fine in Windows virtual machines so Parallels or Virtual PC running Windows under OSX will work on a MAC.

Alan

@starwarsjunkie - interesting questions. As many are participating in answering the questions, i'm wondering if you could answer a question or two from me. Would you have any advice on how to add the information that you're asking about to the product page? We're so used to knowing the power of the product that sometimes we overlook the fact that it's so different from everything else - perhaps these features are not broadcasted correctly. Suggestions welcome!

For reference, here's the link to the product page: https://www.youtube.com/watch?t=19&v=FGJlRqZ_KRM

There's also a video on the page as well. Thanks in advance and welcome to the website!

Hi DJ I spent about one hour on your web site before asking the question. I saw the video you mentioned. In that video, there is a segment showing programming, but no where is it mentioned that a full time tether to a pc is required for operation. Forget videos. Most technical people assume it is marketing fluff. Put up a link on your home page titled "What is EZ Robot? " When I click on the link, show me clear stick figure drawings showing a basic setup. It comprises of a pc sending wireless commands to the kit, and the kit sending back telemetry to the pc. The pc processes the return data according to the users program, and issues new commands. That makes the minimum requirements clear. Every first time visitor will assume, like I did, that this is a autonomous kit like Arduino. I work for the firm that designs on one of the processors in your kit. I got doubts because I know that processor cannot do vision processing etc. Someone who does not have that knowledge will assume this is a stand alone kit.

Great feedback - thanks for the detailed response. I know that you will be overly impressed with your ezrobot's capabilities! Looking forward to seeing your creations.

Have the same problem, bought the product under the assumption I can do the testing/work on the PC but in the end can load the code to the ezrobot for standalone operation (simular like this is working for the NAO). Going thru all the tutorials could not find the "upload to robot" button. Hence started to look in the "FAQ's". Maybe you could add a capability that certain functions can be uploaded? E.g. if you create a script that only uses servo positions (or other light weight functions) an option will appear to export it?

Edit: I famous Dutch soccer player stated: "Every advantage has it's disadvantage", I guess it is true over here as well. I love the integration to Msft cognitive services, integration of Dynamixel, etc. So I probably need to put a NUC in my robot as well to make it autonomous.