Moving post made under this request.
I open this topic because /e/ dev team plan to develop an assistant so maybe if anyone find an opensource assistant to fork, that would save time, energy and money.
I found an Open Source Voice Assistant called Mycroft. Be careful, the name is very confusing when you read it fast do not read Microsoft
Here is the gitHub repo where you can find the core and more: https://github.com/MycroftAI
I think Firefox is working on some similar project. The name is open voice or something llike that.
Isn’t Common Voice for voice donation to help train AIs? Mycroft seems to be more of a Siri, Alexa, OK Google competitor.
Both are great initiatives.
I dont know if it suitable on a android os, but you have snips :
I hate all these voice assistants. if you are in the train or bus or cafe’ it is worst if everyone is talking to his phone. or if you are working in.a big office with several people which are all talking to his desktop PC or mobile phone you can’t work. Every are too much noises.
Please don’t waste time for a voice assistant. I think there are more and better things to develop.
I don’t use voice assistant neither but as you noticed, people like to use them. /e/ wants to be accessible to everyone without privacy matter. People will be free to remove the voice assistant if they don’t want it.
Off topic : I totally understand your opinion. As you can’t change people, I kindly suggest you to spend your time with people who are in accordance with your values and work with a company suitable your values. I know it’s hard but I could do it and I’m just a normal human. It’s always easy to blame others but if we want a change, we have to change ourselves first. We can continue this conversation in private for more details if you want. Cheers
We need voice assistants that aren’t tied to information data warehouses like Google and Amazon. Just let me control my device without making me give up all my privacy and data please.
I don’t like these ‘voice assistants’. I like technology to work as technology. Even a talking GPS gives me the creeps.
This title says it all: a lot of people just don’t like technology to work as technology. So they 'll use voice assistants anyway. Though it’s a good thing to work on privacy respecting voice assistants. That said, none of the existing voice assistants work in my native language, so there is really still a lot of work to do
Maybe you could participate in this project: https://voice.mozilla.org/
The problem they are facing is, after the software is ready for it, to find people to speak in the complete ISBN 90 6648 427 6, all three parts of it, ISBN 90 6648 428 4, ISBN 90 6648 429 2, ISBN 90 6648 430 6, and after that is done to find all kinds of natural voices. It is really a very very big work. The reason why even companies as Microsoft, Apple, Google, Amazon & co only can make it really work in the most important world languages. Know what you start if you want to create an Open source Voice Assistant. Literally thousands of volunteers you’ll need to achieve a goal like that. But maybe I’ll talk to Mozilla now and then.
Mycroft have instructions for building for Android but I’ve never tried them so don’t if they work. AFAIK to actually use Mycroft you need to pay a subscription as the “intelligent” bit is in the cloud.
Mozilla Common voice is not a project to make a voice assistant, it has two other objectives.
- Gather as diverse and complete a voice dataset as possible that is free and open for all to use to train their voice recognition software. They started with English, they’ve recently added French and others.
- Train their DeepSpeech voice recognition engine to make a free and open speech2text engine (AFAIK).
However, Mozilla are working on a voiced controlled browser, which can probably work as an assistant.
Mycroft and Mozilla are working together on these things - Mycroft’s recordings can be used to train DeepSpeech (opt in).
Thank you for your useful information
IMNSHO, any voice assistant needs to be an add-on package and not integral to the system. I suspect a lot of people interested in this project are also interested in no voice assistant at all. I certainly am. I’m not opposed to the development of a voice assistant, I’m simply not interested in using one. Just a few thoughts.
I agree with you.
But as far as I know, this is what /e/ team in planning all additional app will be removable.
i am of the opinion that we have to consider our handicapped fellow human beings, and therefore the speech input should be realized with an urgent priority in all cases.
translated from german to english with the deepL translator.
As a side note mycoft use the Mozilla open voice project I believe to generate the voice for mycroft mark11 home assistant.
I agree this project would work well, it’s just done a successful funding drive and is taking on a lot of staff. Would be a dynamic project to collaborate with.