Skip to main content

Home Screen

Directly after logging in, you'll arrive at the AI-Public home screen.

AI-Public Dashboard

You'll see several elements here:

  • At the top, the number of chat messages and images for the current month (depending on your license)
  • In the left center, the latest news, updates, or helpful tips
  • Bottom left, the images you have generated (if applicable)
  • Bottom center, the files you can chat with (if uploaded)
  • In the right column, possibly Roles, Templates, and Links (depending on configuration)
  • At the bottom, the prompt (question) bar

The Prompt​

At the bottom of the screen is a large input field. Here you can ask a question, or the 'Prompt'. Many articles, videos, and even books have been written about "Prompt Engineering". This refers to formulating the right question to get the best possible answer from the language model.

Getting a good answer depends on the chosen language model and the prompt.

The chosen language model: AI-Public offers different language models that perform better or worse for specific tasks. Roughly, you can rank them from less good to better: GPT 3.5 -> Gemini Pro -> Claude 3 -> Claude 4.0 -> Gemini 2.5 Pro -> GPT-4.1. The most advanced models like Claude 4.0, Gemini 2.5 Pro, and GPT-4.1 all perform very well but also consume the most resources. The choice depends on the complexity of your task and the required accuracy.

The prompt: learning how to prompt is a very important skill in the modern work environment. If you learn to guide the language model well, you can get much better and more relevant results for your work.

Continuing the Chat​

You can then continue chatting with the language model. The message history is maintained. This allows you to 'continue' chatting without having to repeat what you said in previous questions, which is efficient for ongoing projects or analyses.

Switching Language Model

Unique feature of AI-Public

Switching Language Model​

AI-Public has a unique feature: the ability to switch between language models. We offer different language models, each with their strengths and weaknesses. This allows you to discover which model works best for your specific tasks. You can also have a favorite model for different applications. You switch to another language model by clicking on the language model selection button to the left of the answer.

Language Model Selection AI-Public

Copying or Regenerating an Answer​

Next to the answer, there are two buttons on the right. With the left button, you can copy the answer. With the right button, you can have the answer regenerated by the same language model. This can be useful if you want to give the model a second chance to deliver a better result.

Generating Images​

Generating images in AI-Public works via the "Create Image" tool. This tool can have different states:

  • On: AI-Public will try to generate an image based on your prompt.
  • Off: No images will be generated, even if your prompt requests one.
  • Auto: AI-Public analyzes your prompt. If the prompt explicitly or implicitly asks for an image, the tool will try to create one.

You can manage the status of the "Create Image" tool and other tools in your preferences or via the tool settings in the interface. When an image is generated, you can typically choose from different text-to-image models, although the selection may vary.

For more detailed information about tool usage and settings, see the article about Tools.

Chat with Documents​

You can chat with documents by clicking the paperclip button 📎 to the right of the question bar at the bottom. You must first choose a suitable language model before you can click on this, such as GPT 4.0 or Gemini 1.5 Pro. You can upload and select different PDF documents. Once the chat function is on, the contents of the documents will be included when answering your questions.

You can read more about document chat in the related article.

Chat with Internet Sources​

You can chat with internet sources by clicking the internet button on the left side of the question bar at the bottom. Once the internet search function is on, a search query will first be performed based on the prompt, and then the results of that search will be included when answering your questions.

You can read more about internet source chat in the related article.

Roles​

AI-Public offers the ability to set roles. A role could be, for example: 'Marketing Content Creator', 'Legal Assistant for Contract Analysis', or 'HR Support for Policy Questions'.

You give the role a name and then set the initial prompt and system prompt that belong to the role.

The initial prompt is literally the first question you ask in the chat. This field can also be left empty.

The system prompt works in the background to guide the AI's behavior, and you can use this to provide useful background information (context). For example, to give instructions about tone-of-voice, specific company guidelines, or to provide background information about a project.

You can read more about roles in the related article.

Tools​

AI-Public has various tools that the AI assistant can use to help you with specific tasks. These tools expand the AI's capabilities with external functions such as:

  • Create Images: Generate illustrations or visuals for presentations and reports.
  • Internet Search: Find current market information, news, or competitive analyses.
  • Image Search: Search for relevant images for projects or documentation.
  • Data Analysis (example): Assist in interpreting datasets (depending on model and integration).
  • Weather Request: Get current weather information for logistics planning or events.
  • AI Content Check: Detect if text is AI-generated, useful for originality checks.

The AI automatically decides when tools are used, but you can also manually adjust these settings in your preferences or specify them per chat.

You can read more about tools in the related article.