Skip to main content
Squeditor natively supports interacting with AI models (like Gemini) to assist users in writing content or generating images. By attaching the ai_config object to an eligible control, you can define the type of generation, provide boilerplate prompts, and limit boundaries. You only need to define ai_config once per control schema.
AI generation is currently only supported by three control types: text, textarea, and media.

​
Text / Textarea

Used to instantly generate headlines or body copy.
squeditor.json
{
  "id": "heading_text",
  "label": "Heading",
  "type": "text",
  "ai_enabled": true,
  "ai_config": {
    "type": "headline",
    "max_length": 40,
    "prompt": "Generate a compelling headline for a SaaS business tool",
    "variations": 3
  }
}

​
Media

Used to generate images based on prompts directly into the Canvas.
squeditor.json
{
  "id": "user_image",
  "label": "Image",
  "type": "media",
  "media_type": "image",
  "ai_enabled": true,
  "ai_config": {
    "type": "image",
    "aspect_ratio": "16:9",
    "prompt": "A modern startup office building...",
    "variations": 4
  }
}