🎄 50% Off - December OnlyLand Your License Before Prices Soar!
Check PricingInitially sortly has a list of all the generic file types for each category e.g pictures, videos, files, audio.
These allow it to quickly identify what base category each file belongs to.
Under files, it also has a brief understanding of e.g programming, video editor (adobe etc) files - this allows it to identify e.g project folders etc.
These are just for identifying files and folders only.
On initial load and based on settings (timed sorting) or per file change sorting - the sorter does a few things from a few angles.
e.g the cron job means the sorter runs on a specific date and time
or the filewatcher will listen to file changes in dirs and start a sort.
Advanced sorting methods powered by artificial intelligence.
Templated Sorting
(normal and offline)
AI Based Sorting
We use AI to understand your folder structure and files
Language Sorting
Blank slate - tell the LLM what you want to happen and it will sort per your instructions - asking you for permission first as well
Each move operation goes into a single place - it checks if this file is even movable safely e.g Disc space, Safety and finally moves, checks if the op is a success.
One of the protections in place is when the file is small, to simply copy it to the new location, check if done, then delete the old copy.
When the file is really huge, it will move without safety (protection algorithms are being developed for this extreme case).
You are able to request for a unique sorting style via Natural Language and the template will be previewed to you to confirm.
Each part of your request gets turned into a template structure indicating where and to folders, folder types or regex matching patterns to move from and to.
this request can talk about file size, file names etc and all these steps sink through an algorithm that matches these files that pass through your rules for them to be moved.
Compressed versions of your images are sent to our hosted vision models - which can tell us what is in the picture with a greater level accuracy. Based on this, it is measured if a new folder is needed or to use any of the current folders.
You can use rules engine to indicate what level of skew you want.
e.g If the picture includes majority flowers, it should go into the Greens folder in folder x.
Continuous improvement and safety features.
These algorithms have been tested on more than 3million+ files and have performed more impressive than manual sorting.
Several pictures that did not seem to belong together were impressively similar in the most impressive ways possible.
But if a bad decision is made, you can revert from the file activity to return the file to its old spot and it will be skipped from future sorting.
Future versions will introduce the ability to exclude any directory from being sorted if you wish.