Inspiration
Our inspiration for this project was the theme of this hackathon: spy. We wanted to build a real intelligence tool that is more than an AI chatbot. A person can find information simply using Google. Our project streamlines this process and creates less work for the user.
What it does
Our project is an agentic inference tool. When given a URL, image, or username, our tool creates a thorough report of that given element.
How we built it
We used Python for the backend to create our agents. We used Flask and HTML/CSS to connect the backend to the frontend. We used an open-source GitHub repository, WhatsMyName, to help scrape usernames. Gemini was our preferred LLM to use for our agents.
Challenges we ran into
One challenge we ran into was scraping a large amount of data. To counter this, we had to parallelize our processing to speed up scraping times. Additionally, calling Gemini was also time-consuming. We had to make trade-offs and understand when and where we can make these calls. Lastly, being able to use all these different tools, such as Gemini, Flask, and HTML/CSS.
Accomplishments that we're proud of
We are proud of being able to create a project using all of these new tools. Using these new tools was a daunting task. We were able to overcome these challenges and utilize each one.
What we learned
We learned how to use APIs and how to scrape data from an image, URL, or username.
What's next for Sigma
Our next step for scaling would be video inputs and real-time processing and reports.
Log in or sign up for Devpost to join the conversation.