⬜ AI image request

The AI image request block allows you to convert textual descriptions into dynamically generated images via integrated Artificial Intelligence models (fal.ai). It transforms narrative prompts or user input into unique visuals, tailored to creative choices or audience interactions.
This block is ideal for enriching interactive experiences with personalized visual elements, such as avatars, scenery or magical objects, without requiring pre-existing graphic resources.

📥 Inputs

• prompt (Text): Detailed textual description of the image to generate (ex: "A cat flying in space in futuristic cyberpunk pixel art style"). The more precise the prompt, the more faithful the result will be to the request.
• negative prompt (Text): Elements to exclude from the generation (ex: "blur, text, deformed limbs").
• model (Choice, default fal-ai/flux-2/klein/4b): The image generation model to use. Available values: fal-ai/flux-2/klein/4b and fal-ai/flux-2/klein/4b/base/edit.
• image (Image - only shown when the selected model is fal-ai/flux-2/klein/4b/base/edit): reference image to edit/transform.

📤 Outputs

• result (Image): The generated image data (URL and metadata), usable directly in an image or background display block.

💡 Example of use

Scenario: Creating a custom fantasy landscape

1. The player is invited to describe a magical place via a Text entry block: "An enchanted forest with trees with golden leaves and a crystal river under a purple sky".
2. The entered text is passed to the prompt input of the AI image request block.
3. The negative prompt is set to "blur, text, watermark" and the model left on its default.
4. The block is activated, and a loading indicator is displayed during generation.
5. The result is retrieved and passed to an Image Display block to show the generated landscape on screen.

⚙️ Technical Details

• Performance and costs: Image generation is a resource-intensive process. Plan for smooth loading states ("Creation in progress...") and monitor AI quotas defined in your creator account.
• Available models: the model dropdown currently offers two fixed values (fal-ai/flux-2/klein/4b and fal-ai/flux-2/klein/4b/base/edit); the latter automatically reveals the image input.

Documented version: v1

Updated on: 06/09/2026