Generate an image and add it to an article

With the image generator you create an image from a text description — optionally enriched with knowledge from your vaults — and then add it to an article in your knowledge base.

Step by step

1. Open the image generator — via the Image Studio menu item.
2. Set the frame — choose a provider and model, plus an aspect ratio and image size. Depending on the model you can supply reference images (up to 4; some models take fewer or none).
3. Describe the image — in the prompt field, write what should be shown.
4. Optional: enhance the prompt — Enhance Prompt enriches the description with knowledge from your vaults; the sources used appear as “Based on: …”.
5. Generate — the image is rendered from the (enhanced) prompt; to skip the enhancement step, use Generate Directly. Refine produces a revised variant from an adjustment instruction. Every result is collected in the gallery.
6. Add it to an article — save the image from the gallery and insert it into the desired knowledge article (open or create an article in the knowledge base and embed the image).

Good to know

• Knowledge enrichment (RAG) only takes effect in the Enhance Prompt step, not in pure rendering. It is role-based, limited to the content visible to you.
• External image model? If you work with a model marked External provider (own key), keep photos of people and pictures of documents containing personal data out of the reference images — PII protection does not check them. Detected names in the prompt are replaced with placeholders such as [PERSON_1]. → Data protection with external image providers
• Gallery — images are stored with provider, model, dimensions and date, and can be searched and deleted.
• Labeling requirement: AI-generated images are subject to labeling under Art. 50(2) of the EU AI Act — observe it at the latest when using them outside the app.

What’s next

• Image generator — all features at a glance
• Art. 50(2) AI image labeling
• Go deeper: RAG & search