SlothSort demo: open-source AI file tidier (Gemma + Ollama)

Description and use

SlothSort uses a local Gemma3:4b model to read the first 1500 characters of files and suggest names and folders. After user approval, files are moved into a "Tidied" subfolder, speeding up organization without internet access.

Problem

The Downloads folder contains many unnamed or poorly named files that are difficult to identify without opening each one, making manual sorting time-consuming.

Outcome

Using SlothSort reduced the time to organize 13 files from 20 minutes 13 seconds manually to 4 minutes 3 seconds total, with the scan taking 2:2 minutes and user interaction 2:1 minutes.

Workflow

• Point SlothSort at a target folder
• The Flask server scans files and reads the first 1,500 characters of text files
• For each text file, the local Gemma3:4b model suggests a folder and a short clear name via a JSON schema
• The UI displays the suggested plan and waits for user approval
• User reviews and edits the plan as needed
• User presses "Tidy selected" to move files into the "Tidied" folder
• The tool records moves in undo.json and supports "Undo last tidy" to revert changes

Evidence

SlothSort does the sorting for me. I point it at a folder and press one button. It reads my documents one by one and suggests a clear name and a folder for each.

Source

Files are never deleted or overwritten, they move into a Tidied folder inside the folder I picked, and Undo last tidy puts everything back.

Source

The AI makes one small decision per file. For PDFs, Word files, .txt and .md, plain Python reads the first 1,500 characters.

Source

It shows progress. A scan of 40 files takes a while on a laptop, so the server sends each result as soon as it is ready and the page adds rows one at a time.

Source

I sorted the same 13-file test folder two ways. By hand: 20 minutes 13 seconds, opening every file to work out what it was. With SlothSort: 4 minutes 3 seconds in total.

Source