2. Structure of the initial input mask
- Select the desired text model for creating the text content of your news record
- Select the desired image model for creating the images of your news record (if desired)
- if applicable, set additional configuration parameters (currently only for Midjourney)
- Select the individual fields of your news record that should be generated
- If available, select a predefined prompt template and insert it by selecting it into the input field (if no prompt templates exist, this option is not available)
- Create your prompt or modify your selected prompt template.
- Generate your news record
Automatic use of the pre-configured language (of the page)
The contents of the content elements are always generated in the respective pre-configured (page) language, regardless of the language in which the prompt is formulated!
If the (page) language and the language set for the respective backend user differ, it may also happen that the backend labels (e.g., in German) and the generated structure (e.g., in English) are displayed in different languages.
After the fields of the content element have been created, you can now validate the results and make appropriate adjustments.
- For RTE fields, the corresponding editor is also available as usual
- within the RTE field, an additional plugin for modifications is available
- For images, the behavior depends on the selected model
- Midjourney
- Preselection of images
- Selection and generation of an image using the buttons
- processed image with title suggestions
- DALL-E
- direct image suggestion with title suggestions
- Midjourney
- All images can be regenerated separately
- Create/save content element or completely regenerate (Start at point 2 with pre-filled values Entry at point 2)