This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady
My downloads folder can get messy pretty quickly, so for the Hacktoberfest Week 3 challenge, I decided to build something that could help me organize it.
I made Sorta, a small AI-powered app that suggests where a downloaded file could belong based on its filename.
What I Built
Sorta is built with Python, Streamlit, Ollama, and Gemma 3 1B.
I enter or select a filename, and Gemma suggests a folder category. I can review the suggestion in the app before deciding what to do with the file.
This is an early prototype: Sorta currently previews folder suggestions. It does not automatically move files, and its undo feature is not implemented yet.
Instead of having to think through a folder category for every download, Sorta gives me a quick suggestion to review. I haven’t measured a specific time saving yet, but I wanted to explore how AI could make this small, repetitive task easier.
Demo
codesnippet-lab
/
Sorta-AI
A local AI-powered downloads organizer that uses Gemma to suggest folders for your files—without sending filenames to a cloud AI.
Sorta — AI Downloads Butler 🪄
Sorta is a file-organization assistant that uses a locally running Gemma model to suggest folders for your files. You can review the suggestions before taking any action.
Features
- Upload multiple files in a Streamlit interface
- Use local Gemma AI through Ollama to suggest folders
- Preview the proposed organization
- Keep the user in control: this prototype does not move, rename, or delete files
Built with
- Python
- Streamlit
- Ollama
- Gemma 3 1B
Run locally
1. Install Ollama
Install Ollama for your operating system: https://ollama.com/download
2. Download Gemma
Open a terminal and run:
ollama pull gemma3:1b
Ensure Ollama is running.
3. Set up Sorta
From the project folder:
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
4. Launch the app
streamlit run app.py
Open the local URL printed in your terminal.
Privacy and safety
Sorta sends filenames to the locally running Gemma model to generate…
Code
Github Repository:https://github.com/codesnippet-lab/Sorta-AI
How I Built It
I used Gemma 3 1B through Ollama, with the model running on my own computer. For this prototype, the filename is sent to the local model rather than a hosted AI API.
Using an open model gave me a chance to experiment with AI locally and connect it to an app I built myself.
Why Does Open Innovation Matter?
As a first-year Computer Science student who’s still learning development, this project helped me practice connecting a local AI model to a Python interface and handling the model’s response.
My next steps are to add safe, user-approved file organization and implement a real undo feature.
Thanks for checking out my project! I’d appreciate feedback and ideas for improving Sorta.
Top comments (0)