Ok, so I am accustomed to Google and Amazon photo services and their search ability to ID objects and facial recognition.

I'm guessing that Picvario is built using one of those engines.

For example, I have a picture of a female lying on the beach. The POV is from just off her hip so you can see from a couple of inches above her knees down to her sandy toes and beyond that, the crashing waves on the beach. The name of the file is: piq-id-zqlhtvu.jpg

Picvario processed the photo with these results:
Title: Sandy feet on the beach
Description: Sandy feet on the beach with waves and horizon in the background.
Tags: female, woman, adult,healthy,legs,young,summer,one,relaxation,open,sea,travel,outdoor,nature,freedom,vacation,water,joy,leisure,casual,beach,shore,waves,blue,lifestyles,holiday,ocean,20-25 year old, horizon, port,cheering,feet,inspire,nails,relax,sea water.
The Exif data is sparse, only including the width, height, filesize, filename, and a bunch of zeros for geolocation.

So, I would say that I'm pleasantly surprised at how well the image recognition is.

A lightbox feature would be nice so that one could save some images and then download them all as a zip. You do have the option to create a new Collection to add images to, and then you can Export -> Download, and it will zip them up for you to download. For my workflow, that will end up having too many collections and cause me to go back and delete them when I'm done. To me, a Lightbox is a bit more transitory, but I can see saving a collection when working with a larger team.

The only other suggestion for a search filter would be the ethnicity of the people in the asset.

Add in the 4TB of AWS storage, and this is a Great Deal.