Animus Sorter

Sort and name your training images — the companion to AnimaForge.

Turn a pile of scraped images into a clean, training-ready dataset in minutes. Open a folder, sort each image into a subject bin with tactile one-key "punch" presets — one photo at a time, like sorting physical prints — and hit Execute. Every binned image is renamed and moved into a sorted/ subfolder, with the filename encoding exactly who's in the picture.

The punch keys work like latching car-radio presets: press a key to arm a bin (it lights amber, the others gray out), then click an image to fly it in — or drag a tile straight into any bin.

THE FILENAME IS THE DATA

Animus Sorter renames files into the convention AnimaForge reads to derive trigger tokens automatically: NAME_SERIAL_CATEGORY.ext

Aria_001_Character.png — one subject

Aria-Garden_004_Character.jpg — two subjects (hyphen separates names)

Picnic basket_012_Object.png — spaces allowed inside a name

Morning-Evening_002_Style.gif — works for any category

Load the sorted/ folder into AnimaForge and your Name Cast pre-fills straight from the filenames — no find-and-replace across hundreds of caption files.

THE WORKFLOW

1. Open a folder and pick the project category — Characters, Objects, or Styles (one per project).

2. Name your bins — one subject set per bin (Aria, or Aria + Garden).

3. Sort — arm a bin and click images in, or drag tiles. One image at a time.

4. Execute — every binned image is renamed NAME_SERIAL_CATEGORY.ext and moved into sorted/.

Fully local. No network, no accounts, no uploads.

GET IT

GitHub: github.com/SillySilk/AnimusSorter

pip install -r requirements.txt

run.bat

Requires Python 3.11+ (Windows). Dependencies: PyQt6, Pillow.

100% free and open source (MIT). Built for the Anima community — pairs with AnimaForge.

------------------------------ ② VERSION v1.0 ------------------------------