Web Search and internet access

[RaulKong898]

Hi qwersyk,

I'm a user of the Newelle app and I'd like to suggest some improvements that I think would significantly increase its usefulness:

Gemini Advanced Access: I want to be able to use Gemini Advanced features directly within Newelle, without needing a separate app.
Internet Access: Newelle should be able to access and process information from the internet, just like a web browser. This would allow it to provide more complete and up-to-date responses.
File Upload: I want to be able to upload any type of file to Newelle (documents, PDFs, spreadsheets, etc.), not just images.
Eliminate API and Google Account: Ideally, these features should be available without needing to configure an API or connect to my Google account.
I believe these improvements would make Newelle a much more powerful and versatile virtual assistant, capable of providing a more complete and efficient user experience.

Thank you for your time!

[FrancescoCaracciolo]

• Internet access: this is not possible using Gemini API, however it can be externally implemented with extensions if someone wants to do it
• File upload support is planned for LLM in general, refer to issue Sending/Receiving files to/from to models #61
• Eliminate API: not really a thing you want to do with Gemini, free API limits are good and updating cookies etc is actually harder for the end user, so we won't provide it (unless a good library implements it easily).

I'm renaming the issue to "internet access" since is the only doable thing that does not intersect with other issues

[RaulKong898]

I have another idea: the program should be able to run commands with 'sudo' and have access to all system commands. Whenever 'sudo' is used and a password is required, it should be possible to type it in to continue running the command. It should have access to all terminal commands because currently it doesn't have access to all of them. Also, please provide an RPM version in addition to the Flatpak version.

[RaulKong898]

It appears you're pointing out that Google's Gemini model, specifically on the ai.google.dev website, has access to the internet and suggesting that this capability should be added to the Gemini API. You mention that this functionality is accessible through the "Grounding" feature when using a VPN to the US or physically located in the US.

[FrancescoCaracciolo]

Please stop generating things with AI here. The second message is complete nonsense.

If the command uses sudo, you can run it in an external terminal clicking the icon in the code block. By default it uses gnome-terminal but you can change it in the settings.
For other packages format, open another issue.

[RaulKong898]

Please stop generating things with AI here. The second message is complete nonsense.

If the command uses sudo, you can run it in an external terminal clicking the icon in the code block. By default it uses gnome-terminal but you can change it in the settings. For other packages format, open another issue.

I'm not very good at English. I also used to translate the text into English.

[FrancescoCaracciolo]

Not exactly what was asked, but I made this extensions that is related to the issue
https://github.com/FrancescoCaracciolo/AI-WebNavigator