One of the problems with search I find is that the feature also seems to scan for the name of the author, causing the results to potentially be bloated with avatars you weren't looking for depending on how vague the request is. It would probably be better implemented as something toggleable and not on by default.
In addition, the system could use an upgrade in how it displays results. Anyone who has used Favcat can attest to its interface being much more useful than the barebones of VRC's limited control. Being able to arbitrarily increase the number of rows that display results and it remembering up to where you had scrolled to in that set while it's expanded would make for a good QoL features.
One of the problems with search I find is that the feature also seems to scan for the name of the author, causing the results to potentially be bloated with avatars you weren't looking for depending on how vague the request is. It would probably be better implemented as something toggleable and not on by default.
In addition, the system could use an upgrade in how it displays results. Anyone who has used Favcat can attest to its interface being much more useful than the barebones of VRC's limited control. Being able to arbitrarily increase the number of rows that display results and it remembering up to where you had scrolled to in that set while it's expanded would make for a good QoL features.