Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add support for integrated Voice Typing #169

Open
aproposnix opened this issue Oct 7, 2020 · 24 comments
Open

Add support for integrated Voice Typing #169

aproposnix opened this issue Oct 7, 2020 · 24 comments

Comments

@aproposnix
Copy link

In Gboard, there is an icon on the top right of a microphone. Clicking on this will start the voice typing within the Gboard. This makes it easy to quickly correct mistakes in text as the keyboard is visible.

With openboard it's different, the microphone icon is displayed, but when you click it, it just opens the google voice typing interface. You then do not have access to the keyboard - you would need to exit the Google talk app.
@Jhon-1
Copy link

Jhon-1 commented Oct 11, 2020

Voice typing should be a very high priority for Openboard. Good open source is the way to go. And no trackers!

@aproposnix
Copy link
Author

... Good open source is the way to go. And no trackers!

Unfortunately, there is no voice to text application in the FOSS world. I know there is Mozilla's Deepspeech but that is not ready for the masses + they cut funding to the project :(

@Jhon-1
Copy link

Jhon-1 commented Oct 14, 2020 via email

This was referenced Feb 23, 2021
Closed
Open
@nshmyrev
Copy link

Unfortunately, there is no voice to text application in the FOSS world.

There is Vosk, try it. Hopefully we'll integrate it soon:

https://github.com/alphacep/vosk-android-demo

@aproposnix
Copy link
Author

Unfortunately, there is no voice to text application in the FOSS world.

There is Vosk, try it. Hopefully we'll integrate it soon:

https://github.com/alphacep/vosk-android-demo

Thanks for the information. I tried out the app, but it doesn't seem very useful as it doesn't integrate with anything. Are you hoping to make openboard the first integration?

@Jhon-1
Copy link

Jhon-1 commented Feb 24, 2021 via email

@jnnkB jnnkB mentioned this issue Mar 16, 2021
Closed
@nshmyrev
Copy link

A project that integrates Vosk with openboard is here:

https://github.com/ccoreilly/LocalSTT

@Jhon-1
Copy link

Jhon-1 commented Mar 31, 2021 via email

@ewheelerinc
Copy link

vosk-api now has a working jitpack build: alphacep/vosk-api#558

and it is available on maven as vosk-android: https://repo1.maven.org/maven2/com/alphacephei/vosk-android/

Perhaps this is easier to integrate than the vosk demo. Would love to have voice typing with microG, so +1 for this feature!

@ewheelerinc
Copy link

ewheelerinc commented Nov 10, 2021 •
edited
Loading

Also https://github.com/Stypox/dicio-android has integrated vosk-android if it helps understand how to integrate the API into openboard:

This was referenced Nov 10, 2021
Open
Open
Open
@ewheelerinc
Copy link

Also as @theoden8 notes: "There's also kaldi, android demo is here. The code doesn't seem long."

One way or another there aught to be a great voice package to plug into Openboard!

@ewheelerinc ewheelerinc mentioned this issue Nov 10, 2021
Open
@LuccoJ
Copy link

LuccoJ commented Nov 10, 2021 •
edited
Loading

There is also Kõnele which is an Android app that provides a speech recognition interface to other apps (maybe now through a separate service that I'm not sure was a thing last time I checked...?), and while it's primarily tested with Estonian, it is possible to use it for English thanks to a GStreamer server for the speech recognition engine Kaldi, installation of which can be simplified thanks to an available Docker image for it.

I tried setting this up on my home server a long while ago and it was relatively easy thanks to the Docker image, and it worked, although it wasn't very good at recognizing my somewhat heavily accented English. I had a video to show it but can't find it anymore... anyway, I came here due to the MicroG mention, and now I see that @ewheelerinc did mention Kaldi just above, so well... throwing this out here, and confirming that it can work for some definition (and if I could get it to work...)! There used to be an English demo server available with the Kaldi GStreamer service but it appears to be down right now. I am not aware of any other public serves.

@ewheelerinc
Copy link

ewheelerinc commented Nov 13, 2021 •
edited
Loading

@LuccoJ It would be neat if Kaldi could be installed locally on the phone and then point Konele at localhost!

@LuccoJ
Copy link

LuccoJ commented Nov 13, 2021

@LuccoJ It would be neat if Kaldi could be installed locally on the phone and then point Konele at localhost!

I think the documentation says it is possible to compile it for Android and run it there; however I have not tried that, so I don't know how to get it to work... Vosk clearly does run on Android, and to be honest, I think when I tried the demo app it recognized my speech better than Kaldi did with the English model.

@karam72
Copy link

karam72 commented Mar 6, 2022

This missing feature is only reason to not install OpenBoard most of my friends. :(

@aproposnix
Copy link
Author

This missing feature is only reason to not install OpenBoard most of my friends. :(

It's definitely the reason I am not using it.

@ildar
Copy link

ildar commented Dec 29, 2022 via email

@tristan-k
Copy link

BTW AnySoftKeyboard can use it while the OpenBoard seems not seeing the service.

How do I use Vosk with OpenBoard? I would prefer to use Dicio. But neither OpenBoard nor AnySoftKeyboard are supporting it. Instead I'm forced to use Google.

@ildar
Copy link

ildar commented Apr 9, 2023 via email

@LuccoJ
Copy link

LuccoJ commented Apr 9, 2023 •
edited
Loading

@ildar that is not entirely accurate. Dicio integrates Vosk, and recently it also started exposing itself as a standard speech recognition API for Android, and as such it can be used as one by other applications (although I think it may be missing some parts since IIRC some apps see Kõnele but not Dicio for me).

So while the speech recognition engine is still Vosk, Dicio is similar to using the Vosk demo app, but potentially with a more advanced interface (although using Vosk through Kõnele is another route).

@Jhon-1
Copy link

Jhon-1 commented Apr 13, 2023 via email

@Mhowser
Copy link

Mhowser commented Apr 14, 2023

@Jhon-1 did your account get hacked?

@Jhon-1
Copy link

Jhon-1 commented Apr 20, 2023 via email

@Redsandro
Copy link

Redsandro commented Aug 21, 2023 •
edited
Loading

So now there is Whisper by OpenAI and it's really nice. Someone ported the "tiny" model to Android and I'm surprised how good it is from running the test app on my 5 year old phone. Discussion here.

Someone else attempted to make a voice keyboard out of it, but I don't know the status - I can't get it to run.

Either way, there are interesting FOSS options now, in case someone is eager to work on this.

Unfortunately, there is no voice to text application in the FOSS world.

Related:

dziaineka pushed a commit to dziaineka/openboard that referenced this issue Nov 19, 2023
@Helium314
set background on each emoji / clipboard start instead of onFinishInf…
f27baef 
…late

fixes openboard-team#171 and apparently openboard-team#169 too
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
10 participants
@ildar @tristan-k @ewheelerinc @LuccoJ @Redsandro @nshmyrev @aproposnix @Mhowser @Jhon-1 @karam72