Aegis Shield

Website home page
Trying a prompt.

Inspiration

On the news, we have seen countless reports of employees putting in sensitive information, including financial details, code infrastructure, and more, into AI, causing data breaches to steal their important information.

What it does

Our project can hide sensitive information and replace it with placeholders so that outside AI will never receive the information.

How we built it

We first built a privacy detection engine that identifies sensitive information, such as credit card numbers, passwords, API keys, and financial data, and replaces it with secure placeholders while preserving the original meaning and context. We then developed a web interface that integrates with Ollama, allowing users to safely interact with AI models while ensuring that confidential information is protected before it is processed."

Challenges we ran into

We were unable to use a Gemini API Key, so we had to rely on Ollama to be able to run the LLM.

Accomplishments that we're proud of

We are proud of drawing the logo and background of the website by ourselves, and being able to create a smooth project that will truly benefit people.