Creator (previously called MAGES) is going through a transition.
Many videos, text, and UIs may still refer to it as MAGES while this rebranding is underway.
Navigate to Personalized AI, through the navbar on the left and click on the Personalized AI Creator.
Here are the steps:
Step 1: Create a personalized Library, in which the agent will be based on.
If your data can be summarized to a few paragraphs you can skip this step.
Step 1: Library Resources.
Fill in the library details.
Here you can select if the library will be visible and editable by only the author, or the whole organization (Pro Feature).
Step 2: Library Files.
Attach the files that will form the library.
Supported files are: pdf and txt.
Creator Pro users can opt to summarize their libraries, which can help agents navigate large amounts of data.
Step 3: Library Creation.
After creation, the library will be available to select in the creation process of future agents.
Now that a library has been created, the next step is to create an agent.
Step 2: Create a personalized agent.
Step 1: Agent Details.
Fill in the agent details.
In the prompt section describe how you want your agent to behave.
If your dataset is not very long you can simply describe it here instead of creating a library.
Step 2: Agent Libraries.
Choose among the libraries of the organization.
The libraries selected are the ones which the agent will be based on.
This step is optional.
Step 3: Agent Creation.
After creation, the agent will be available to all the users of the organization as well.
The agent ID will be available to be used in a unity agent as well.
Step 3: Interact with the agent.
There are two ways to interact with the agent:
Portal Interaction: In the AI assistant page, choose the agent from the dropdown and begin to chat with it.
Unity Interaction: Set up an embodiment agent inside Unity with the ID of the agent provided. Inside Unity the agent can be interacted with via voice, not text.