Hi everyone,
I’m using Make.com to send image generation requests to Leonardo.Ai, and I’d like to clarify a specific use case:
I want to upload an existing product photo to Leonardo, and get back a new image that reflects both my prompt and the uploaded image as visual reference.
My current setup:
- The image is uploaded via the
UploadInitImagemodule in Make - It is successfully stored in my Leonardo media library (I can see it listed)
init_image_id: set as{{35.id}}(from the upload step)init_strength: set to0.8- Prompt includes:
“Use the uploaded image as reference. Preserve layout, shape and style.”
What works:
- Image is accepted with no errors
- Result seems loosely influenced by the reference
- Prompt affects style and mood as expected
But in Leonardo Web UI:
- The generated image does not show “Image Guidance” as active
- The reference image is not displayed in the generation history preview
- It behaves like a pure text-to-image generation
- So it’s unclear if the reference image is actually used, or just ignored
What I’m trying to achieve:
I want Leonardo to treat the uploaded image as a true visual base, just like it does when uploading an image manually in the Leonardo web interface (Image2Image mode).
Ideally:
- The output image would preserve object structure and position
- The UI would show the reference image as “used” in the generation
My question:
- Is this possible via API?
- Or does Leonardo only show “Image Guidance” as active when using the manual upload on the website?
- Has anyone achieved this behavior via Make.com or direct API?
Thanks in advance!