Welcome to the Question2Answer Q&A. There's also a demo if you just want to try it out.
+10 votes
1.5k views
in Plugins by
edited by

Free Download - GitHub

Look Ma! No Hands!!

Voice Commander is a Narrow AI plugin that allows you to Search, Navigate and interact with your Q2A website trough voice commands.


This plugin uses HTML5 Speech Recognition and Annyang, in order to convert matching words into voice commands.

The plugin is fully translatable, with the ability to add customized responses.
Check out this list of supported languages.


Navigation 

Use commands such as: Search for [something], Go to homepage, Go to questions, Go to tags, etc to open main navigation links. 

While viewing question lists, you can say: Open the nth question (1st/2nd/...), which is pretty much self explanatory.


Interactivity

You can create/ask questions using only your voice: Set title, Set body, Set tags. Once you're happy with your input, you can say: Send or Submit.

You can also use commands such as: Scroll up/down, Scroll to answers, Answer question, Comment or Do comment, and much more...


As of now, Voice Commander is only fully compatible with Google Chrome, and websites that use HTTPS protocols.

To avoid missing functionalities from partially supported browsers, a message will appear for those browsers, suggesting to use Chrome.

Free Download - GitHub 

by 1 flag
Which AI is needed and how much that costs?
by
@polle No AI or extra stuff is needed. The plugin uses HTML5 Speech Recognition
https://developer.mozilla.org/en-US/docs/Web/API/SpeechRecognition
which handles voice data, checks for matches, and processes them into commands/tasks.
That's why this is "Narrow AI" instead of "General AI".

I might convert this into a free plugin in a couple of months, since I'm stepping down from active Q2A development, so time management will be focused for Polaris theme, whenever necessary, so I won't have a million things to work on whenever there's crucial updates to the Q2A platform.

If you want to get the plugin right now and support my work, I can make it half the price it is right now. Let me know.
https://rxchun.github.io/shop/voice-commander/?ref=q2a-voice-commander-post

(All people who've purchased the plugin so far will be mentioned as supporters of this project, in this question.)
by
Hey, thank you for explaining this is not AI and it is a voice commands plugins.
by
No worries @polle . Here's a picture that further describes the diferences between 'Narrow AI' vs 'General AI' in simpler terms.
https://i.ibb.co/hZ3MKbv/01-03-02-general-vs-narrow.jpg

Sourced from the interaction-design.org article:
https://www.interaction-design.org/literature/topics/general-ai#narrow_ai_vs_general_ai:_what%E2%80%99s_the_difference?-1

2 Answers

–1 vote
by
Yes great but it is very expensive man
by
+1
Expensive? Did the price change? Seems remarkably inexpensive.
0 votes
by

This plugin is now free!

A few days ago while I was working on some updates for the plugin before releasing, the plugin was working just fine. Although today I tested it, and sometimes the `speechSynthesis` works, sometimes it doesn't.

It doesn't throw any errors to the console, but I believe it's due to a deprecated action that before was allowed to autoplay sounds and speech, but now requires a user interaction with the website first, before you're able to use it. More specifically this patch:

`speechSynthesis.speak()` without user activation is deprecated and will be removed.

I did patch that for the sound autoplay, a year or so ago, for when a command is recognized. But now it seems they're pushing the same thing for the speechSynthesis as well.

So be aware that the plugin may not always work consistently for all your users, for now.
If someone is willing to fix it and push a commit, you're more than welcome. To fix the issue you'll have to apply the same logic I did on #this section, but for the speechSynthesis now.

Might as well remove the localStorage saved information, previously implemented so that it would autostart when a user returns to the website, as now it will only work with user interaction first.

...