 OK, so hello everyone. Yesterday I will talk a bit about KXGPT and its date. KXGPT itself is a relatively new project, so I think we started one year at this time ago. Who have you ever heard about KXGPT? OK, some of you, perfect. Hopefully I'll get some new information for you. So first things first, why did we start building KXGPT? At first, because AI was cool at this time. So we thought we have to do something with the AI and thought this would be a good use case. The first thing we found out is that travel shooting, kiwanitas can be really hard. So the people who started with KXGPT are all susubmins, old SREs, and so on. And we all had some issues with kiwanitas. We travel-shooted day by day and so on. And we thought this should get easier. And this brings me to the point we are travel-shooting the same issues over and over. I think many of us deal with missing service accounts. We all deal with labor on pods, which are not there for services and so on. And we thought this could make it easier. We also thought that we as humans have the problems, but we also have the ability to find them. And the AI itself is able to provide solutions for our problems. So for instance, if we collect enough information to find out all of the related resources and so on, we could put this into an AI model and could say, yes, this is the problem. So what is KXGPT? At first, KXGPT initially was only a CLI. The first command we had was KXGPT Analyze. And with KXGPT Analyze, no AI was in place. The second thing we had was some kind of explain mode, which sent some data to OpenAI and so on and got information for it. So the KXGPT CLI consists of some analyzers, such as a port analyzer, deployment analyzer, and so on, which tries to find problem patterns. Furthermore, we have some kind of AI integration. And in the meanwhile, we have many AI integrations, such as for OpenAI, Vertex, Bedrock, and also for LocalAI. And the interesting part is with the analyzers, we try to find as many data as possible in our Kubernetes cluster. And with the AI integration, we try to interpret it. So in the end, the KXGPT tells you that you have a problem, and this is the solution. Without the AI integration, it only tells you that you have a problem. So a typical result of the KXGPT CLI would look like this. So in fact, this was a problem that image could not be pulled, so I tried to find the first screen that I found today. And it tells you that you're in a port with the name test star in the default namespace. You have the problem that image cannot be pulled. And with the details, and this is the AI generated part, you get the information what you can do with it. So as I said before, we started one year ago. And in this year, many things happened. At first, I think not completely a year ago, we released the very, very first version of KXGPT. And found out that many people liked the project. And also, we found out that many people would like to contribute. So after some days, we got the first 1,000 GitHub stars. I think in the meanwhile, we have 5,000. And thought it would be nice to get on the Zinsev landscape, and therefore, we had to change the logo twice. So at first, because our logo was a bit too fancy for SVG, and the second one because we didn't like our colors. Good. After some time, we found out that the CLI is pretty nice, but we could automate some things around it. So we started creating an operator, tried to automate some things, and tried to automate all of this. So in the meanwhile, you can run KXGPT as an operator, and it detects problems automatically when they are cured. After, I think, three months, we sent over the sandbox proposal to the Zinsev. We were a pretty young project. Therefore, it took a bit longer. And in December, we got accepted in the sandbox. In the meanwhile, and since then, we've worked on lots of features and integrations, and the project grows day by day. So what's planned? We refactored the whole integration part of KXGPT. We will add more integrations, such as for AWS, other cloud providers, and other operators. We will enable the scanning of CRDs. We will have additional outputs at the moment. You can send Slack messages. In the future, it will be teams and so on. And obviously, we will try to integrate with other AI mail tools. So one of our maintainers also was the initiator of the AI working group. So this would not be possible without a community. So for everyone who tried out KXGPT until now, and for everyone who contributed to KXGPT, a big thank you at this point. And if you'd like to get in touch with us, we have a GitHub repository, obviously. So please star us. Please contribute, raise pull requests, write issues, and so on. Share your thoughts. So we also have a Slack channel. Tell us what you would like to see in KXGPT. And if you feel comfortable, feel free to raise a pull request. And if you'd like to talk to some of us, we are in the Project Pavilion on Wednesday in the afternoon. With this, I want to thank you. And have a nice day here at KXGPT. Thank you.