Project VUBA

AI Assisted Linux Terminal

Comment

Inspiration

What inspired me the most was the absence of an AI-supported terminal to help beginners learn Linux more efficiently. Additionally, I got inspiration from Cortana in Windows and Siri in iPhones. I wanted to create a similar tool for Linux, particularly tailored for ethical hacking purposes.

What it does

The system receives chat messages from users, spanning from casual conversation to executing commands. It then generates responses similar to a chatbot. For instance, if a user sends a message like "start a Django project called happy," the system executes it within the built-in terminal django-admin startproject happy and responds with a confirmation, such as "Done."

How I built it

The frontend was constructed using ElectronJS, and I integrated the terminal into the application using technologies like Node-pty, XtermJS, and websockets. For the API, I primarily used Flask due to its simplicity. Regarding the AI functionality, I fine-tuned an OpenAI model to generate responses matching my intended style and commands.

Challenges I ran into

My initial experience with ElectronJS was baffling, but I eventually grasped its workings. Fine-tuning the model presented another hurdle initially; however, with persistent efforts, I managed to achieve satisfactory results.

Accomplishments that I am proud of

I take pride in the entire project, especially in successfully implementing the AI model, crafting an aesthetically pleasing UI, and ensuring the functionality of the command execution feature.

What I learned

This project taught me how to provide more effective prompts to AI and create robust datasets. I also enhanced my skills in JavaScript, honed time management abilities due to it being a solo project, and gained considerable knowledge through the build-and-learn process.

What's next for Project VUBA

For Project VUBA's future, I'm considering two paths: one is to make it an open-source project where the community can contribute with improvements and fixes. The other option is to turn it into a commercial product.

Share this project:

Updates