[FREE] Voice Recognition extension without Google Dialogue

The voice recognition process continues until the system recognizes the first phrase. Further, it is assumed that the user will process the received information and start the recognition service again.

1 Like

Great extension!
But I noticed some problems:

  1. StopASR - StopASR does not stop listening immediately.
  2. Clicking Stop and then restarting returns an error frequently (this did not happen in version 6):
1 Like

I didn’t have such a problem. Does this error occur in the compiled and installed application too?

1 Like

There is an error in your plugin. Without google dialog, the codes work in campanion but not in apk app. This error occurs only when google dialog is closed, please fix it

2 Likes

If you think this problem is caused by my application, don’t worry, I tried it in the applications I created from scratch, but the problem was still there.

1 Like

I tested it on different devices - the apk works. What version of android API do you have?

1 Like

I’m using Android version 11 but I don’t think this will happen because I found the apk of the old first version of my same application and tested it. It only works in campanion.

i am trying to make an artificial intelligence based assistant and this plugin is really important to me

1 Like

Version 11 (API 30) may have new security requirements. I have not tested the work of my extension on this version of Android. Companion differs in that it gets all possible system permissions during installation.

hmm ok, I tried giving all permissions and it didn’t work :slight_smile:

The problem is this:
For apps targeting Android 11 (API level 30) interaction with a speech recognition service requires element to be added to the manifest file:

<queries>
   <intent>
     <action
        android:name="android.speech.RecognitionService" />
   </intent>
 </queries>
 

I don’t know yet how to do this from an extension. Perhaps more experienced developers can tell me.

1 Like

Could the coders team help with this? :slight_smile:

I reported it to a developer from Turkey, I’m waiting for a response, thanks for your interest.

2 Likes

can you give the code of your plugin please

I see 3 solutions:

  1. Decompile the APK of your application - edit the manifest file - compile the APK again. Look here and here.
  2. Ask the Kodular developers to add the annotation we need.
  3. Master Android Studio and create applications without limits ))
2 Likes

Well, You can modify manifest using rush

The sources of the published extensions are not secret, as they are in the .aix file itself and can be easily extracted. But, in this case, the source will not help.

Yes, maybe this is a solution. Need to try. I think that the developers of the APP Inventor will solve this problem, because the Package visibility requirement applies to many already developed extensions.

1 Like

Kudos on creating an amazing and easy-to-use extension!

I’m a language teacher looking for a solution where:

  1. I can define a list of keywords to be noticed by the speech recognizer
  2. Speech Rec. can continuously listen
  3. Each time a word from the list is recognized, the score is incremented by 1.

Basically, I’m looking for a tool to count how many of the pre-defined keywords the user-learner has used in a 1-5 minute speech.

Any idea how I could realize this?

Oh sure. Such an application is not difficult to create using a Kodular:

  1. One button for saying keywords. All recognized words must be saved to the list. This list can be saved to a file so that it does not disappear when the application is restarted. Also provide a button to clear the list.
  2. Another button to start recognition and counting keywords in the text. Each time the student stops speaking, the application counts the matches of words from the list of keywords with the recognized text. Also increases the total counter (or counters per word). And so on until the teacher stops the recognition process. Counters are reset when starting a new recognition.
    It’s all!
1 Like