FiloSort AI: Building an AI-Powered File Organization Assistant 🤖📁

#devchallenge #hf26challenge #ai #webdev

Have you ever spent too much time searching through a cluttered Downloads folder, trying to remember where you saved a particular file?

That problem inspired me to build FiloSort AI, an AI-assisted file organization project developed for the Hacktoberfest 2026 DEV Challenge.

The idea

FiloSort AI aims to make file organization simpler by helping users identify suitable categories and suggesting where their files could belong.
Instead of manually deciding how to organize every file, users can explore organization suggestions through a dedicated web interface.

What I've built

Dashboard: An overview of the application's organization features.
File categories: Separate sections for documents, images, audio, and video.
AI Organiser: An interface for generating and reviewing file organization suggestions.
Suggested organization: Recommendations for categories and destinations based on file information.
Backend API: A FastAPI backend that processes requests from the frontend.
Deployed web application: A live frontend and backend that make the current prototype accessible online.

Tech stack

Frontend: React, Vite, CSS
Backend: Python, FastAPI
AI experimentation: Ollama with Gemma 3 (1B) during local development
Deployment: Render
Version control: Git and GitHub

Demo

Try the live application:
Live Demo:-- https://filosort-ai-1.onrender.com
Backend API:-- https://filosort-ai.onrender.com/docs

FiloSort AI is a web-based prototype that helps users explore file categories and organization suggestions through a dedicated interface.

Code

GitHub Repository: https://github.com/aditric7/FiloSort-AI
The project uses a React frontend and a Python FastAPI backend. The repository contains the source code for the application.

How I Built It?

I built FiloSort AI to explore how AI-assisted organization could make managing digital files easier.
The application includes a dashboard, file-category sections, an AI Organiser interface, and a backend endpoint for generating organization suggestions.
During local development, I experimented with Gemma 3 (1B) through Ollama. One important lesson was that a model running locally on my computer is not automatically accessible to a cloud deployment. The current deployed version uses fallback suggestions when the local model is unavailable.
At this stage, the application focuses on organization suggestions rather than automatically moving files on a user's computer. Connecting hosted model inference and supporting actual file-content analysis are planned improvements.

Why Does Open Innovation Matter?

Open innovation made this project a valuable learning experience because I could explore open-weight AI models, experiment with local inference, and build on accessible development frameworks.
Using Gemma through Ollama gave me an opportunity to understand how a model can be integrated into an application without relying entirely on a closed, hosted AI API during local development.
Open-source tools such as React, FastAPI, and Ollama also helped me learn by building and debugging a complete application, from its interface to its backend and deployment.
For me, open innovation is about being able to experiment, inspect how tools fit together, learn from the community, and improve a project incrementally.

DEV Community