Our inspiration for this project was based on a real need to review packet capture in a timely manner, in a way that is even remotely human. What Wirelight essentially does, is that it acts a lot like wireshark in the sense that it can capture live traffic at whatever network it is pointed to, providing a detailed analysis report about the traffic. Where we differ drastically, is that we have developed LLM integration to allow an AI of your choosing to read the traffic data, be it from a live session or from a packet capture file sourced from other means. When it came to building it, we started out using Claude to build the groundwork of the project, providing us with a clean, professional UI dashboard and the basics of AI integration. Eventually, we migrated the project over to ChatGPT for debugging, fixing issues and implementing the rest of the features, which saved us plenty of time. Relatively speaking, we ran into very little issues throughout the project, yet nothing felt like it wasn't a challenge to overcome; the hardest part about it all was getting everybody into the same VPN so that our computers could communicate with each other and with the test homelab throughout development. In the end, we were very happy with how our demonstration video came out, and we like how it can employ Nmap to learn your network before querying the AI with the packet captures as well as any other prompts. Having a VPN to fall back on as a group with consistent IPs was a very important lesson to learn, because it makes collaborating as a group during these events far easier. We learned video editing skills, prompting skills and networking skills when it comes to locally hosting AIs and communicating with them through the local network, making use of API keys and we re-learned GIT and made a public repository. Following this event, it'll probably go into private use among the creators of the project, to point it at our own networks for a variety of purposes, to help understand and clean up our own networks.