Explore, Learn and Master at Topqlearn

What is Image to Image Feature in Leonardo.ai

Leonardo.ai’s image-to-image feature is an incredible tool that allows you to turn your wildest creative visions into gorgeous visual art.

By providing Leonardo with a reference image and text prompts, it can generate new images for you while retaining the angle, positioning, and posture of objects in your original image.

I’ve used this feature extensively in my work as a digital artist and have been blown away by what it can produce.

In this comprehensive guide, I’ll walk you through exactly what image-to-image is, how to use it, tips to get the best results, and awesome examples of what it can do.

Get ready to spark your imagination like never before!

What Exactly Is Image to Image in Leonardo.ai?

Leonardo’s image-to-image feature falls under the broader category of AI image generation. Essentially, it leverages the power of artificial intelligence to generate brand new images based on source images provided by the user.

The key defining capability of image-to-image is that it replicates the core aspects of your reference image – the positioning of subjects, perspective, framing, etc. – while allowing you to provide text prompts to modify, combine, or enhance that image through descriptions.

For example, you could:

  • Take a photo of a bird and ask Leonardo to make it look like a watercolor painting, all while retaining the original positioning/framing
  • Provide a picture of a room and ask to change the wall color, floor materials, furniture style etc.
  • Have a person imaged in various outfits or backgrounds while their pose stays consistent

The common thread is that Leonardo locks in the spatial relationships of the original image as the foundation, then makes changes using the text prompts.

This sets it apart from general AI image generation which works solely from text prompts without a reference image to ground the positioning.

Step-by-Step Guide to Using Image to Image

Ready to start creating? Using image-to-image is simple. Just follow these steps:

  1. Sign up/login to your Leonardo.ai account
  2. Go to AI Image Generation in the left sidebar menu
  3. Scroll down and click on Image Guidance
  4. Upload your desired reference image
  5. Adjust Strength – this controls how closely the output matches your reference image’s positioning and framing. 0.35 = no influence, 1 = full replication. 0.60 is a good starting point.
  6. Select your preferred Leonardo Style – Illustration, Realism etc.
  7. Pick an appropriate Fine-Tuned Model
  8. Enter Text Prompts describing what changes you want made or elements to add
  9. Click Generate and watch the magic happen!

Key Settings for Great Results

Beyond just following the basic steps, dialing in these settings will really optimize your images:

  • Prompt Magic – Turn this on. Image Prompts require Prompt Magic V2.
  • Init Strength – As noted above, start with 0.35 then adjust up or down. More replication vs. more text prompt influence.
  • Guidance Scale – Lower values around 5-9 generally work best for tighter control of text prompts.
  • Fine-Tuned Model – Take time picking one that matches desired style. The model has a big impact.
  • Text Prompts – Clear, concise descriptions separated by commas generally works best.

Awesome Image-to-Image Examples

To give you an idea of what’s possible, check out these real images I created using the image-to-image feature:

Original Reference Image

Image to Image Feature

With text prompt: “watercolor painting”

watercolor painting image to image

With text prompt: “watercolor painting, flower”

watercolor painting image to image #2

The way Leonardo.ai instantly understands exactly what I want and transforms images with pinpoint accuracy blows my mind! As you can see, the possibilities here for digitally manipulating photos are truly endless.

FAQs About Leonardo’s Image to Image Capability

Can I provide multiple reference images?

No, you’re limited to one source image per generation. Leonardo needs that single fixed reference point to understand positioning.

Is text input required or can I omit prompts entirely?

You can leave off text prompts if desired and simply generate variations on the original image!

What’s the benefit over DALL-E or Midjourney?

The core advantage lies in retaining precise object positioning. Those other AI’s don’t provide that spatial grounding.

What content works best?

Image-to-image excels at replicating photographs of people, animals, objects etc. and making modifications like above. Fully invented scenes/compositions from scratch can be harder.

Does it work for videos?

Unfortunately not – image-to-image is currently only for static images. But who knows what future capabilities will bring!

Go Unleash Your Creativity!

I hope this guide gives you a comprehensive overview of how Leonardo.ai’s game-changing image-to-image feature works and how you can leverage it for your creative projects.

The ability to take existing images and transform them into visually stunning new pieces of art is incredibly powerful.

The key is learning to balance reference image positioning with text prompt additions to achieve your desired outcomes. With some practice, you’ll be astonounded at what you can create.

So what are you waiting for? Sign up today, supply some images, get imaginative with text prompts, and let your creativity run wild!

The world of AI-powered image generation is at your fingertips.


The Image to Image feature in Leonardo.ai signifies a significant advancement in image manipulation and generation. Its ability to transform images across styles, seasons, and qualities offers immense potential in creative design, photography, and immersive experiences. However, while promising, leveraging this feature effectively demands expertise and consideration of its limitations.

Ultimately, Leonardo.ai’s Image-to-Image feature stands as a testament to the innovation and potential of AI in reshaping visual content and unlocking new possibilities in creative expression and visual storytelling.

4.8/5 - (6 Vote By people)

Last modified: March 7, 2024

Join us telegram channel

Leave a Comment