I've been looking at Dragon Naturally Speaking to add more options to EZ-AI. The question I can't seem to find an answer to is if the ability to program against the DNS API is included in the Professional version of DNS 13 or if I would have to buy the SDK?

The pros to buying the SDK that I see are Example code Documentation Ability to redistribute the DNS dll's with EZ-AI meaning that others who use EZ-AI wouldn't need to have DNS installed

The con is that the SDK is $5000.00.

If I could use the professional version Pros The cost is about $700 Cons are No example code No documentation Anyone who uses these features would need to also purchase DNS and I don't know if it would have to be the pro version or not at this point.

I am adding a chat bot like ALICE to EZ-AI. DNS will make it much better to use and it would be on your network so it would be much faster and your own.

I will be adding the ability to pass text that is spoken to a variable in ARC. This could be used to do anything in ARC and would complement the speech engine in ARC.

I will be adding the ability to browse web sites and search for data on the Internet about spoken words.

More things will come based on requests.

The purchase of the SDK would require me to start charging for EZ-AI to help cover the cost of the SDK. I would be willing to trade 3d printing or other things that will allow me to continue building robots. The $5K would eat into my funds that I have put back to buy 3D printers and controllers, servos, etc. it would however allow me to distribute this work far easier.

I am looking for the level of interest that there is in this project so that I know which direction I should take this. If you are interested in using EZ-AI, and would be interested in trading for this app, or would be willing to drop say $50.00 to $100 on it, please post. I appreciate it.

@ d.cocharn. i thougth you were out for the weekend. i'm in for the 3d parts we talked about last week and in to help cover some costs. i down loaded the patch for the star wars game but get errors on opening the database for adding questions and answer in the database. i know i could not afford this alone so hopefully others will help.

I just got done with my massage. I'm out and didn't bring my computer but the iPad is here:).

I'll chech the script for pushing the questions Monday night. I let another program script that part instead of doing that part of the scripting myself. I'll write it and get it to you on Monday.

I would certainly be willing to pay between 50 and 100 to be able to use DNS with my robots. Pretty sure you need the SDK for what you are looking at doing.

Alan

Maybe a kickstarter would be a good way to raise the money. $100 is a bargain for good AI. We need 50 people who will pay $100 or 100 people willing to pay $50.

I got the chatbot working in EZ-AI. It will take a lot to duplicate on someone elses machine so I need to figure out how I am going to get that done, but it is working.

I also have the text that is coming back from the chat bot being spoken. I will figure out how to do an sayEZB through the application before releasing it.

I also am trying to use the System.SpeechRecognition object to try to understand what is being said and pass that into the application. This is being a bit of a pain and I need to spend some more time understanding the grammer objects. I choose the system object because it is my understanding that it uses the normal grammer objects from the windows operating system. I had hoped that this would allow a far larger speech library to use, but it is acting like it isn't loading the OS grammer objects. I may end up going with the Microsoft.SpeechRecognition object because it is easier to deploy with your application but I would need to figure out how to load a good grammer library.

I am still considering the DNS Developer kit. I just find it hard to drop the money on it. Also, I will see if the google API is available or if it is only for android.

On the android front, I just spoke to my son about how they are programming their app (using an Android and speech recognition to query a database). They have just started working on it and probably wont be done until June. I will probably wait until they learn from their programming efforts before I jump on it. One cool thing that they are planning on doing is using the GPU on the android device for the query functions. If they pull it off, that would be the route I would be going down. This will keep the CPU on the android available for normal functions.

Does that mean we will not be able to use the speech script louis wrote?

You can use that. I will pass it back as a variable to ARC so that this can be used the louis' script. I will probably add a setting in the configuration allowing the user to specify if it should come back as a variable or send it back via the ezb say or just the say command.