Generating images with AI

What is this about?

With the AI image generator you create images directly in ALLMEDIA.AI, without an external tool. You describe the subject in a chat, choose model and format and take the finished image into the Gallery. To make the images fit your brand, you can provide context such as your visual identity, brand voice or target audience.

Brand AI is not a separate image generator. Brand AI analyzes your website and creates a brand profile. You can give the visual identity derived from it (brand colors, fonts and imagery) to the image generator as context.

Prerequisites

• At least the Editor role.
• Enough credits. Each generation uses credits.

Step by step

1. Open the image generator. Go to Content → Gallery, click AI Studio at the top right and choose AI Image.
2. Choose a model. Under AI Model, choose an image model. Models from OpenAI (GPT Image), Google (Nano Banana) and FLUX are available. A notice shows whether the model supports Refining possible or creates a completely new image with every prompt.
3. Define the format. Under Format & channel, choose an aspect ratio or one of the Channel presets, for example Instagram Story (9:16), LinkedIn Post (1:1) or YouTube Thumbnail (16:9).
4. Add context (optional). Via Context or @ in the input field you give the generator additional information: Visual identity, Brand voice, Target audience, Channel, Knowledge store or a Media item from the Gallery.
5. Describe the subject and send. Describe subject, mood and setting and click Send. The generation continues in the background, even if you close the chat.
6. Refine the image. With models that support Refining possible, you describe in the next prompt what should change. The last image is then used as a reference. You can also attach your own image as a reference.
7. Take the image. Click Use image, give it a title, choose a folder and confirm with Save to media library. The image is saved with AI disclosure.

Options at a glance

• Previous generations: Earlier chats with their images are kept. You can still take an image you didn’t use at the time later on.
• Changing the model: If you change the model, no chat history is sent, only your next prompt and, if the new model supports it, the last image as a reference.
• From within the post: In the create area, too, you can create an AI image while writing a post and attach it directly.
• Automatically per post: If a channel should automatically get an image for every new post, use a media pipeline. See Media pipelines for automatic images.

Limits

• One image is created per prompt.
• Which models can use a reference image varies. If a model doesn’t use a reference image, the generator points this out and ignores it.
• If credits are missing, the generation stops with a notice and offers a link to top up.

Frequently asked questions

Do I need an external AI tool or my own API key?

No. The image models in the AI image generator work without your own access. The costs are covered by your credits.

Does the AI learn my brand style through its own training?

No, there is no training of custom image models. You bring in your brand style via context, for example via Visual identity, a reference image or a precise description in the prompt.

Where do I find my generated images?

Images you have taken over are in the Gallery in the chosen folder, with AI disclosure.

Related articles

• AI Studio: AI images, AI videos and carousels
• Brand AI: the fast setup of your system
• Managing the Gallery (media library)
• Media pipelines for automatic images