AI Images allow you to generate images using text prompts directly within your templates. Instead of providing a URL to an existing image, you can describe what you want, and RenderForm will generate it for you using OpenAI's gpt-image-1.5 model.
This feature is available exclusively for RenderForm Pro users.
How it works
To generate an AI image, use the ai: prefix in the src property of any image component.
The text after the prefix will be used as a prompt to generate the image.
{
"template": "<TEMPLATE_ID>",
"data": {
"my-image.src": "ai:a cute bunny with colorful Easter eggs on a green meadow"
}
}
RenderForm will:
- Detect the
ai:prefix in the image source - Send the prompt to OpenAI's image generation API
- Replace the value with the generated image
- Render the final template with the AI-generated image
Full API example
curl --request POST \
--url https://get.renderform.io/api/v2/render \
--header 'X-API-KEY: <API_KEY>' \
--header 'Content-Type: application/json' \
--data '{
"template": "<TEMPLATE_ID>",
"data": {
"title.text": "Product of the Day",
"background.src": "ai:minimalistic abstract gradient background in pastel blue and pink tones",
"product-photo.src": "https://example.com/product.jpg"
}
}'
You can mix AI-generated images with regular image URLs and text properties in the same request.
Image specifications
| Parameter | Value |
|---|---|
| Model | gpt-image-1.5 |
| Resolution | 1024×1024 pixels |
| Quality | High |
| Format | PNG |
Pricing
Each AI image generation adds 20 credits to the request cost. This is in addition to the base render cost.
For example, if your template uses 2 AI-generated images, the total additional cost will be 40 credits (2 × 20 credits) on top of the standard render cost.
Tips for writing prompts
- Be descriptive — the more details you provide, the better the result. Include colors, style, mood, and composition.
- Specify the style — mention if you want a photo, illustration, watercolor, 3D render, flat design, etc.
- Keep it concise — while details help, avoid overly complex prompts. Focus on the key elements.
- Avoid text in images — AI models often struggle with rendering text accurately in images. Use RenderForm text components instead.
Good prompt examples
ai:professional product photo of a red sneaker on a white background with soft shadows
ai:flat illustration of a mountain landscape at sunset with purple and orange sky
ai:minimalistic geometric pattern in navy blue and gold colors
Limitations
- AI image generation adds processing time to the render request (typically 5–15 seconds per image)
- Generated images are always 1024×1024 pixels and will be scaled to fit the image component in your template
- Prompts must follow OpenAI's content policy. Prompts that violate these guidelines will be rejected
- The
ai:prefix is only supported for thesrcproperty of image components