SlothSort AI — Local AI-Powered File Organizer

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady 🦥 SlothSort AI: A Local AI File Organizer That Works at Your Pace What I Built I built SlothSort AI, a local AI-powered file organizer that helps keep folders tidy without manually sorting files one by one. SlothSort uses a locally running AI model to classify files by their metadata and suggest categories such as Documents, Images, Code, Archives, Videos, Audio, and Other. It creates a preview of the proposed changes, so users can review the destination and approve the organization before files are moved. The idea is simple: let AI handle the repetitive suggestions while keeping the human in control. Demo Screenshots: The project README includes screenshots of the dashboard, AI-generated organization preview, and organization result. GitHub repository: https://github.com/Shubhamkumar-op/SlothSort SlothSort currently runs locally. A live hosted demo is not available yet because it depends on a local Ollama installation and model. Code Explore the source code here: 👉 SlothSort AI on GitHub The project is organized into separate modules for scanning files, generating AI suggestions, validating the organization plan, and executing approved file operations. It also includes automated tests for important scanner and executor behaviors. How I Built It I built SlothSort using Python and Streamlit, with Ollama running Gemma 3 locally for AI-powered classification. The main components are: Python: Core application logic and file handling. Streamlit: Interactive dashboard for scanning files and reviewing suggestions. Ollama + Gemma 3: Local inference to generate file organization suggestions. Pytest: Automated tests for scanning, validation, and safety-related behavior. Git and GitHub: Version control and open-source project hosting. The workflow has four main stages: Scan eligible files and collect metadata. Ask the local AI model to suggest a category, filename, and destination. Display the proposed organization plan for human review. Validate the approved plan and move eligible files within the supported demo folders. SlothSort uses file metadata rather than uploading file contents to a cloud AI service. File movement requires explicit user approval, and the application checks for unsafe paths and destination conflicts. Why Does Open Innovation Matter? Open innovation makes it possible to build useful AI tools without depending entirely on proprietary APIs. Using an open-weight model through Ollama lets SlothSort run inference locally, giving developers more control over their environment and reducing dependence on external AI services. It also makes experimentation more accessible: developers can inspect the application code, modify the workflow, test different models, and contribute improvements. For a tool that works with personal files, local inference is especially valuable because file metadata can be processed without sending it to a hosted AI API. My Agent Session I developed SlothSort through an iterative workflow: implementing the scanner, connecting local inference, adding preview and approval steps, and testing file-operation safeguards. I don't have a shareable DevRelay session link to include yet, so I'm leaving this section out for now. The Lazy Payoff SlothSort is designed to reduce repetitive file sorting by generating organization suggestions automatically instead of requiring users to decide where each file belongs manually. I tested the workflow with a set of demo files and used the dashboard to review and organize them into category folders. I haven't measured a reliable time-saving figure yet, so I won't claim a specific number. The goal is to make file organization a quick review-and-approve task rather than another chore. Prize Categories Select the partner prize categories that match the challenge rules and the features actually demonstrated by SlothSort. I haven't selected a category here because the applicable prize options need to be confirmed from the challenge page. Built with Python, Streamlit, Ollama, and Gemma 3. 🦥