We suggest that you start your creation journey by exploring the creation of Text AIs, but if you are already familiar, let's learn how to build an image AI with your style.
Welcome to the fascinating world of Tess AI, the leading AI platform in Latin America. Here, users can create and publish their own AIs in seconds! All of this is connected to the largest AI platforms, such as ChatGPT, LLaMA (from Meta), Bard (from Google), Dall-E, and MidJourney.
There are infinite possibilities. Logo creation, realistic images, sales emails, advertising campaigns, and mainly what we will address in this article: the possibility of creating your own image AI!
Creating an Idea
The prompt is the central part of your agent. It will dictate what will be created. Therefore, we will begin with it.
Let’s say you have the idea of creating images using a “Steampunk” style.
To do this, we will make a simple test command for the AI, trying to create a dog. Therefore, the prompt will look like this:
“a dog Steampunk style and high detail.”
For better results with image prompts, always opt for creating the prompts in English.
Testing and Validating the Prompt
With your initial prompt in hand, just test your idea with the help of Tess. To do this, select “For Image” in the “AI Copilot” menu.
After that, select the Tess image model you want to test and insert the prompt you developed.
The prompt in this case was satisfactory, so we will proceed with the same one. Thus, we will use it in our AI.
“**element** Steampunk style and high detail.”
Note that where "a dog" was previously, we replaced it with "**element**." This will be our variable, and we must use it in the construction of the AI.
AI Studio
Now that we are aware that the prompt is functional, it is time to bring a model to life. Thus, the first step is to access your model area in the “AI Studio” menu.
These initial steps are identical for creating both text and image templates. But we will recap them here.
- Click “Add Agent” and start creating your custom model.
1. Creating the Agent for Images
For the initial filling: we will begin by defining the type of AI (chat, image, text). The process is simple, and all you need to do is choose from the options, and today, we will choose the Image option. If you wish, you can switch between AI models, filtering by a specific model.
After that, we will move on to the more technical part of the creation process.
2. Prompt
Idea: An agent that creates images in the Steampunk style. We want the AI to create images in different contexts but in the same style.
Tested prompt: "**element** Steampunk style and high detail."
3. User Inputs
It is at this stage that we create the fields that users must fill in. In the case of the prompt in question, these are the "elements" that must be present in the image.
Notice that we specified in the prompt, and now we must create the input, which we will do below. Select the “User Inputs” option, choose Short Text, and name it however you prefer (in our case, we will name it "Element").
Remember that you can test other types of inputs in your template; just select the one that best suits your project.
4. Finalizing the Agent
Here is the final step of creating your AI model, where you have already structured the prompt and inserted your User Input in it.
Observe that if you do not use the inputs between asterisks in the Prompt, nothing that the user types in the input will be used by the AI. Therefore, make sure it is filled out exactly as in the created inputs, including the **asterisks**, or click on “+” to fill it out automatically.
With that done, just save the model and test it! Your agent will be stored in AI Studio, and you can publish it if you wish.
Review
After creating the model, if you wish to publish the agent, it will undergo a review by the Pareto team, who, upon not identifying any discrepancies, will approve it for use. But you can use and test it while it is in approval. Just click on the link provided in AI Studio.
Contents
For more details on how to create prompts for generating images, we recommend watching the webinar “Creating Amazing Images with Tess AI,” conducted by Rica Barros, CEO of Pareto. Don’t miss this opportunity to deepen your knowledge!