braden-w / whispering

https://whispering.bradenwong.com/
MIT License
261 stars 28 forks source link

Development Will Resume Soon! #72

Open braden-w opened 8 months ago

braden-w commented 8 months ago

Hello all, just wanted to say that I have not forgotten this project! I've been checking in from time to time really surprised and grateful for the growth and feedback!

You may have seen my commit from about two months ago, and I just wanted to echo this in GitHub issues in case someone hasn't seen it in the readme!

I've been busy with university in the past half year, but plan on resuming development to address all the issues at around February. Hopefully, this coincides with Svelte 5's release: it's coming soon, and I would prefer for my updates to be written in the new syntax rather than old syntax if possible!

DavidGP commented 8 months ago

That's very awesome Braden, looking forward to it! After all, Whispering is by far the best speech recognition on Windows out there.

CognitiveTwo commented 7 months ago

That's great, this is a really cool project!

OLH21 commented 7 months ago

great project we are waiting for you to come back to it. thanks for your commitment it is usefull to a lot of us

Arche151 commented 7 months ago

It's great to hear that, @braden-w ! Thanks a lot for the update :)

Quick question: Would it be possible for you to release a hotfix, so that the Firefox version of Whispering works again? Because the only reason, I'm still using Brave Browser is, that Whispering is broken on Firefox, unfortunately.

And another question: Could you kindly explain why Whispering requires access to all sites data? That's kind of concerning for me as a privacy advocate.

DavidGP commented 6 months ago

Hi Braden, just wanted to mention that we users are thinking a lot about you and would love to see you resume development of Whispering. For my part, I would like to have two different hotkeys to turn Whispering on and off so that I can control it reliably with Autohotkey and/or keyboard remapping.

OLH21 commented 6 months ago

Hi Braden, just wanted to mention that we users are thinking a lot about you and would love to see you resume development of Whispering. For my part, I would like to have two different hotkeys to turn Whispering on and off so that I can control it reliably with Autohotkey and/or keyboard remapping.

Having two hot key is a mess for normal user. I have another reccorder that have that and it’s such a problem. Let Whispering with simplicity of use. It’s perfect as it is nothing more needed.

DavidGP commented 6 months ago

Understand what you mean and agree. The problem that I'm having is that the status of Whispering (if minimized) is not clear, Therefore it can be active and record for hours without you noticing, which can also lead to unnecessary costs with OpenAI. If you then want to actually dictate something and switch it on via hotkey, it will be switched off instead and you loose your dictation.

Therefore, an option for two different hotkeys (that can but does not have to be used) would be good, alternatively a way to recognize visually and programmatically (particularly when minimized) whether Whispering is active and listening or not.

[dictated using Whispering]

braden-w commented 6 months ago

Dear all, thank you for the responses and I appreciate the interest! I have just returned have actually been doing some informal development and trial and error under the dev branch.

I expect to have a pretty big rewrite of a lot of the application logic, given that I have learned a lot as a developer in the time being.

May I ask, for all issues mentioned in this thread—could you guys create a new issue for each feature request? Once I finish this rewrite, I will go through each of the issues mentioned and attempt to address them in subsequent releases.

Thank you all once again for the continued support!!

DavidGP commented 5 months ago

Thank you Braden! Looking forward to seeing the next version of Whispering, which is already one of the greatest things in existence for dictating anywhere on your PC.

May I ask, when resuming work on Whispering, could you start building from this great fork by @HamzaETTH? https://github.com/braden-w/whispering/issues/69 Thank you!

Arche151 commented 4 months ago

@braden-w The extension is completely broken now unfortunately, due to the domain for ChatGPT moving from chat.openai.com to chatgpt.com.

Do you think you could kindly provide a hotfix?

doxgt commented 4 months ago

Therefore, an option for two different hotkeys (that can but does not have to be used) would be good, alternatively a way to recognize visually and programmatically (particularly when minimized) whether Whispering is active and listening or not.

[dictated using Whispering]

Most definitely. Alternatively, if we are sticking to the single-key-toggle, please consider enabling some sort of distinguishing feature in AutomationProperties Class for the respective buttons (other than the obvious visual difference), so as to facilitate better accessibility, on Windows at least.

doxgt commented 4 months ago

Greetings everyone. While awaiting Braden to hopefully resume development of this wonderful, cross-platform app with browser plugin, take a look at my comment below:

https://github.com/braden-w/whispering/issues/51#issuecomment-2119316375

With vanilla AHK (v1 or v2), cURL , and phiola, it wouldn't take much to whip up a Whispering desktop app equivalent, with prompting.

That I have done myself.

Of course, such an app would be Windows-only, and extra work would be involved to make a browser plugin. But if you go through clipboard, which the browser would see (as long as proper permission is set), a plugin may not be needed.

AHK is not cross-platform, of course. But the resulting product would still be portable from 1 Windows rig to another.

There will be enormous flexibility with the cURL approach on Windows. The workflow can even be ported to something like VoiceMacro, for both computer control and prose dictation ...

P.S. fwiw, I cobbled together the following:

https://github.com/doxgt/PlayGround/blob/main/Simplest_cURL_GPT_API_Test.ahk

https://github.com/doxgt/PlayGround/blob/main/GPT_cURL.ahk (requires phiola as a portable voice recorder controlled through CLI)