Deploying Large Language Models (LLMs) has never been easier thanks to the EOSC EU Node Tools Hub and your personal User Space. With streamlined integration and automated deployment workflows, users can now launch powerful models inside their sealed user environments like Mistral in just a few clicks using Ollama. This approach simplifies setup, reduces operational overhead, and ensures a smooth path from experimentation to production-ready AI services.
Disclaimer: This LLM tool is offered via the Tools Hub “as is” without any warranty of any kind. It runs inside the personal User Space and deployed over the user’s personal cloud environment. The European Commission and service operators assume no responsibility or liability for any errors, service interruptions, data loss, or damages arising from its use. The tool can be used exclusively for non-commercial, research, educational, and experimental purposes.
Let’s see a guided tour
Visit the EOSC EU Node portal and log in to your User Space using your home institution’s credentials, by clicking on “Log In” in the top right corner.
Once logged in, an overview of your User Space will be shown.
In the left menu, you can see the different services offered by the EOSC EU Node, as well as your current number of available credits. From this menu, select the Tools Hub to see the list of available tools.
In the list, you will see the name of each tool, its description and what type of tool it is: Virtual Machine or Container. Find your desired tool and then verify its type. In this case, let’s select the Ollama + Open WebUI tool. This tool offers a comprehensive, open-source AI stack for research purposes. Ollama provides a lightweight runtime for running and managing large language models. This combination enables researchers to deploy, experiment with, and use advanced AI tools seamlessly within the EOSC EU Node.
Next, you need to allocate an environment of the correct type to host the tool. Again, in the left menu, select the associated service, “Virtual Machines” or “Cloud Container Platform,” and create an environment with the necessary resources for the selected tool. Consider that the amount of resources set in any tool are some default values that can be modified to fit your processing requirements. In this case, the selected tool is designed to utilise the Virtual Machines service, so let’s allocate a Medium-sized environment.
Once the size of the environment is selected, you must select the period during which this environment will last. This form will display the total cost, and the credits that will be deducted from your wallet. Remember that later, you can extend the lifetime, or release the environment early. In both cases, the amount of credits will be adjusted accordingly.
Once the order is submitted, the system will promptly create the requested environment, in one of the available Virtual Machine providers. Once it is ready, you will receive a notification and be able to find the new environment in the Allocated environments section.
Now that you have allocated your Virtual Machine environment, you can proceed with the deployment of the tool. Go back to the Tools Hub section, select the tool and click the Deploy button. In the form, you must select the project name where the environment has been allocated (usually your Default Personal Project or a Group project for collaborative work). It will display the amount of resources required by the tool and the current available resources in your environment, allowing you to determine if there are enough resources to perform the deployment.
In this form, you can also customise the input values and change the computational resources assigned to fit your processing needs. For advanced users, it also enables setting the Ollama models that will be configured in the tool. Once complete, click “Save and Select Project”.
Finally, click “Proceed” to start the deployment.
To view your deployed tools, go to the Tools Hub dashboard, and in the rightmost tab, select the Deployments list. Wait about 10 minutes for your Ollama + Open WebUI tool to be fully deployed and configured. You can check the status of the tool in the bottom left corner; you must wait until it reaches Deployed status to use it.
To access your tool once it is deployed, click on the map icon to see the output values. These values will show you different ways to access your recently deployed resources. In the case of this selected tool, it displays the URL to access the OpenWebUI interface, as well as the IP address and SSH credentials if direct access to the Virtual Machine is required.
Now you can access your OpenWebUI + Ollama tool using the provided output value “openwebui_url”. In this tool, the first step requires creating an admin account; you only need to provide a name, email, and password.
After that, you can access the prompt to ask your questions to the selected AI model. By default, the “llama3” model is selected. However, you can change this in the top left menu, using other European-based alternatives, such as the EuroLLM model generated by a European-funded project (https://eurollm.io/) or the Mistral created by Mistral AI SAS, a French company.
Then, you can ask your questions to the model, and in a few seconds, you will receive AI-generated answers.
Once you have finished your interactions with the AI model and no longer need the tool, you can return to the Tools Hub dashboard, navigate to the Deployments tab, select your tool, and click the trash icon. After confirming the deletion, all the tool-related resources will be deleted.
If you no longer need the created environment, you can release it to avoid using more credits than needed. Go to the Virtual Machines service, select the environment, and click on the three dots in the bottom-right corner. Then, click “Release”.
The release environment form will show you the number of credits that will be returned. Click on the Release button to confirm, and in a few seconds, the environment will be deleted, and any remaining credits will be returned to your account.
With the Tools Hub in the User Space, deploying tools and experimenting with LLMs becomes a seamless part of the research workflow. From selecting the right environment to accessing advanced models in a few clicks, the EOSC EU Node offers a reliable, user-friendly pathway from testing to real-world application. Whether you are exploring Mistral, EuroLLM, or other emerging models, the Tools Hub empowers you to work directly with cutting-edge technologies, securely, efficiently, and within the European research ecosystem. Stay tuned as more tools and capabilities are continuously added, opening new possibilities for research across disciplines.
