Ozeki AI Server Quick Start Guide
To start using Ozeki AI sever the best option is to go trough the steps presented on this page. This page gives you the download URL you can use to downoad the product, it shows you where you can download AI models from, and gives an insight on how to setup your AI models once Ozeki AI server is intalled on your computer.
Step 1 - Download Ozeki AI Server
To start you need to download Ozeki AI Server from the following URL:
The following video shows the download procedure.
To download Ozeki AI Server, you can use the download links on this page or you can click on the Download link in the top menu bar of this website (https://ozeki.chat) (Figure 1).
The download link will open the download page. On this page, you need to select the Ozeki AI server product, and you should click on the blue Download button (Figure 2).
This will take you to the available versions. On this page you should select the latest version and click on the blue link starting with the text OzekiAIServer (Figure 3).
Step 2 - Install Ozeki AI Server
After Ozeki AI Server is downloaded, you will find the installation package in your Downloads folder. The following 1 minute video takes you through the installation steps. In short the installation is similar to the Installation of a standard Windows application.
The installation starts by finding the downloaded Ozeki AI Server zip file in your Download folder. Next you need to extract this zip file using the "Extract All" context menu item in Windows explorer. (Figure 4)
Once the installation is complete, you can run the OzekiAIServer exe file to start the installation wizard (Figure 5). Note that the version number on your system will look different as Ozeki is continuously improves the product.
Once the installation is started you need to read and accept the license agreement then proceed with the standard installation steps. If you are unsure, it is a good practice to simply click Next on each form to complete the installation procedure (Figure 6).
Once the product is installed you will be presented with the final form of the installation wizar, that allows you to Run Ozeki AI Server. On this form simply click Finish (figure 7).
After the installation is complete, you can open the graphical user interface of Ozeki AI Server by typing the following URL into your browser:
On the graphical user interface you need to click on the AI icon to open the AI app, that allows yout to setup AI models and AI chat bots (Figure 8).
Some common tasks we recommend after installation are:
- Change the admin password to allow remote access.
- If you own a license you should activate your license code
Step 3 - Download an AI Model
Once Ozeki AI Server is installed, the next step is to download an AI model. You can download AI models from the huggingface.co website, where over a million models are available. Check out the following video to see how you can download an AI model.
To start the model download procedure you need to navigate to Huggingface:
If you open Huggingface, you will notice that there are AI models in different format. You should select an AI model in the GGUF format. GGUF stands for GPT-Generated Unified Format. To find GGUF models, simply click the GGUF format filter on the Models page (Figure 9).
Once the GGUF model filter is set, you can search the model that best fits your requirements. If you are new to AI, it is best to start with LLama-3.1, which is a great Large Language Model (LLM). LLama has different versions. Each version is described by the model size. When you download LLama you need to pick a model size that will fit your AI hardware (your computer).
The most important parameter of your AI hardware is the amount of GPU memory it has. This parameter determines which model is optimal for you. The following table gives you an idea on which model to download.
| GPU | Model size | Download |
| No GPU, or GPU with less then 12GB VRAM | 1B | Llama-3.2-1B-Instruct-GGUF |
| 1 GPU with 12GB VRAM | 8B | Llama-3.1-8B-Instruct-GGUF |
| 2 GPUs with 2x24GB VRAM | 70B | Llama-3.1-Nemotron-70B-Instruct-HF-GGUF |
If you are unsure, start with the smallest model. Select the 1 billion (1B) model size. Smaller models are faster and have a better chance to fit into your computer's video memory (GPU VRAM). If you have a very powerful system with lots of GPU RAM, select a larger model. Once you have decided which models to use you should save the model file to C:\AIModels on your computer (Figure 10).
C:\AIModels
Save the downloaded model to your C:\AIModels directory.
To learn more about AI model selection and downloads, click on the following button:
+36 52 299 415