Maybe just a collection of examples that have been successful would be helpful here? Requested via slack.

(I'm not sure if this is the appropriate place for feedback like this; please point me in the right direction if not!)

I just took a look at the pricing model for Sidekick. One thing that stood out to me is that the user plans don't seem well-suited to individuals like myself who use BN extensively during shorter periods, and not in a full-time work capacity. E.g. in my case, I tend to have some months of programming, followed by a month or two of reverse engineering, and so on.

I would have to frequently subscribe and then unsubscribe. And fair enough - maybe there's not really a good fix for that, it just is what it is. But it occurs to me that I'm probably not the only person in this boat; I would imagine that many personal/non-commercial users are similar. Such users would never be able to take advantage of the yearly discount, unless they can afford to pay for periods where they're not using BN as much or at all (which I'd argue defeats the discount anyway?).

I think the yearly discount makes a ton of sense for businesses, but it seems to me to not be a great fit for individuals. 🤔 But obviously I don't have any insight into your customer base, so this is all just my perspective/speculation.

Finally, the upfront cost for a year of Sidekick is also pretty steep for an individual. If you stick with the yearly subscription for individuals, I would suggest at least offering a model similar to Google Workspace where you can commit to a full year in advance, but still pay monthly or quarterly (or similar).

Regardless, speaking as someone who's otherwise wholly uninterested in e.g. GitHub Copilot, Sidekick looks like a very cool application of AI and I'm looking forward to giving it a try!

In versions prior to Sidekick 1.1.131, the Sidekick plugin fails to load after an update or install due to the following error:

[Default] A module that was compiled using NumPy 1.x cannot be run in [Default] NumPy 2.0.0 as it may crash. To support both 1.x and 2.x [Default] versions of NumPy, modules must be compiled with NumPy 2.0. [Default] Some module may need to rebuild instead e.g. with 'pybind11>=2.12'. [Default] If you are a user of the module, the easiest solution will be to [Default] downgrade to 'numpy<2' or try to upgrade the affected module. [Default] We expect that some modules will need time to support NumPy 2.

Example:

If I ask Sidekick to add a comment explaining the selected code, what happens is:

If instead I try this code:

I get:

Which seems to indicate that it is describing the code above.

I get similarly odd results anywhere I've tried this feature. Seems like it could use some love in general.

Version and Platform (required):

• Binary Ninja Personal: latest
• OS: Microsotf Windows 11 Home
• OS Version: 10.0.22631 Build 22631
• CPU Architecture: x64

Bug Description:
This is the error I am getting in the logs during installation:

Failed to import python plugin: official/Vector35_Sidekick: huggingface-hub>=0.23.0,<1.0 is required for a normal functioning of this module, but found huggingface-hub==0.22.2.
Try: `pip install transformers -U` or `pip install -e '.[dev]'` if you're working with git main
Failed to load plugin: Vector35_Sidekick

I tried the recommended solutions to no avail.

Steps To Reproduce:
Please provide all steps required to reproduce the behavior:

1. Install Sidekick

*Expected Behavior:
Proper installation of Sidekick into the Binary Ninja platform.

Screenshots/Video Recording:

Binary:
N/A

Additional Information:
I did recently install a lot of plugins, but I also tried a fresh full uninstall and reinstall and received the same problem. However, the plugins didn't get removed upon removal of Binary Ninja from the host after reboot and I believe that the GEF plugin I use leverages the huggingface library which is unable to be installed to support Sidekick.