EDIT: Thanks Dev! You're right.

Only used a little . The smaller icon option is good

REQUEST - that it can be used for PDF documents displayed in Firefox.

Works pretty well (a few errors sometimes though XD)
Looking forward to support for more languages!

it's become very useful and an integral part of my daily usage. the only reason i didn't give it five stars is because i need a dark theme!!!

Promising additions and great initiatives from Mozilla

quite decent but the button really needs to disappear when fullscreen video is playing atleast in minimal mode, its extremely distracting and blocks content. This is the only real issue i have with it but it makes it quite unusable

Great at what it's designed to do, quite useful for catching up on news. One minor hunch; could you please make it so that the orb disappears when going full screen in YouTube? Simple feature that would go a long way. (Maybe a shortcut to hide and show the orb?) Thanks!

I will be following the project with interest. Surprisingly, there is no repository for this extension. Also, no communication platform has been created. For example, a discord server.
The extension is obviously designed for Firefox. Why does its design seem so foreign to the browser's design? Isn't this part of the same ecosystem? I agree with all the comments about the support for different languages, the dark mode, and the annoying extension icon.

Por favor hagan que la IA se pueda comunicar en otros idiomas, en español por ejemplo.

If assuming that you approach AI not as a "think for me" and rather a "think with me" piece of software, this is great. Works with everything Ive tried so far on it.

But only have 2 simple requests.

*Allow DarkReader / Implement a dark mode.*

Its a major eyesore that this is seemingly the only element I cant change.

*Add another option to engage it via the extension icon itself*

This may just be in response to the first request...but still, what can be done in minimal can be done from the icon itself. I dont need this element always stuck on the right side (why cant it be on the left) of my screen.

It struggles significantly with describing entertainment videos. Mixtral makes many mistakes and overlooks important moments, as if it doesn't understand the sequence of the narrative and what to write. The descriptions are too concise. It would be beneficial to recognize when more details are needed and when they are not. In any case, I hope this can be improved, as the concept itself is great. It would also be nice to consider using Llama 3, GPT-4o Mini, or Claude 3 instead of Mixtral7B.

Add Russian language support, please.

it's overall a good extension and ai works well but the only thing i don't like about this is that i have to keep the floating icon, either dynamic or minimal.

Firefox, please add a button next to the address bar, that would be very non-distracting. i know i can pin extension to toolbar, but it still needs improvement and the floating icon will still say.

Other than that, it's really great!

please add Russian language support.

Mozilla offers us this new Firefox extension which provides a free and privacy-preserving AI assistant for web browsing. The quality of the web page and video summaries is very decent IMO although there's room for improvement. Latency is also fine. Regarding the UI, I like the simple and functional design, but find it too invasive by default, while it's fine when set to minimized.
For a beta, this already got me hyped and looking forward for future versions!

Some suggestions for improvement:
* Response quality can be improved, by incorporating more recent and powerful language models than the current choice (Mistral 7B).
* Allow to save a history of summaries/chats.
* Implement natural language search across website pages. That could be super useful for overcoming the lack of a search function on a website.
* For even less clutter, move the Orb button as an menu in Firefox's extension toolbar.
* We could expect more transparency regarding how it works: what exactly is send to the model (does it only see the current HTML page? what is the full prompt?), where is the server located, or what is the exact language model used (if open).
* Add a dark mode option