Canvas
Last updated
Last updated
Ideogram Canvas is an infinite creative board designed for generating, editing, and organizing images. It’s the ideal workspace for projects involving multiple images, whether they’re generated directly within Ideogram or uploaded from your computer.
With Canvas, you can:
Upload your own images by clicking Upload, using copy-paste or simply dragging in an image.
Edit images using two new powerful tools:
Magic Fill (inpainting) to add or modify elements within an image.
Extend (outpainting) to expand your image beyond its original edges.
Generate images using text prompts.
Remix images and compositions.
Download your final composition/work as an image.
You can access Canvas in different ways:
Create a new canvas: Click the Canvas icon on the left-side panel.
For Generate: Select the aspect ratio you want for your images.
For Remix: Position and resize the generation window over the area you wish to remix.
For Magic Fill: Use the masking tools to select areas you want to change, then click Next to adjust the position and size of the generation window. Ensure enough of the underlying image is included so the AI has sufficient context.
For Extend: Position the generation window over the area you wish to expand, including a significant part of the underlying image to maintain visual continuity. Use the masking tools to select areas you want to alter.
Generated images will appear on the canvas.
For Generate and Remix: Four images will be placed directly on the canvas.
For Magic Fill and Extend: One image will be generated where you placed the generation window. Click on the image and use the arrows in the menu below it to toggle between the four generated options.
Located on the left side of Canvas, this panel includes:
Home icon: Returns to the Ideogram home page.
Generative tools:
Generate: Create images using text prompts.
Magic Fill: Inpainting to add or modify elements within an image.
Extend: Outpainting to expand your image beyond its original edges.
Remix: Generate images based on a parent image and prompt.
Utility tools:
Select: Select, move, resize, and arrange images.
Hand: Pan and navigate around the canvas.
Upload: Import images from your computer.
Download: Crop and export your final composition.
This panel allows you to:
Undo and Redo actions.
Zoom in and Zoom out of the canvas.
You can zoom using the + and − buttons or select a predefined zoom percentage from the dropdown menu.
Selecting any image on the canvas will display image actions below it. The options will vary depending on the type of image selected.
When you select an uploaded or pasted image, the following panel appears:
From this panel, you can:
Duplicate the image.
Remove the image from the canvas.
Access More options:
Bring to front
Send to back
Duplicate
Remove from canvas
When you select a generated image, the following panel appears:
From this panel, you can:
Switch between the four generated images.
Open the Right Panel (see below)
Access More options:
Download the image in different file formats.
Bring to front
Send to back
Duplicate
Remove from canvas
Accessed via the menu under any image frame by clicking the right-side panel icon, this panel displays details about the selected image, such as the prompt, size, style used, etc.
Each generative function offers specific and also common options accessible by clicking the Options icon. The common function are explained below while the specific ones will be detailed in the function's section.
Private Generation: Generate privately or publicly.
Aspect Ratio: Define the format and size of the image
Style: Determine the style of the image
Magic Prompt: Enhance your prompt.
Color: Select a color palette for your image.
Select: Select, move, resize, and arrange images.
Hand: Pan and navigate around the canvas.
Rectangular mask: Create a mask from a freeform shape
Freeform mask: Create a mask from a freeform shape.
Brush mask: Create a mask from a freeform shape.
Load previous mask: Loads the last mask you have used.
If you’re a new user, or simply new to Canvas, consider experimenting with the Generate and Remix functions outside of Canvas to familiarize yourself with their features. Also, get acquainted with features like Private Generations, Aspect Ratios, Style, Magic Prompt and Color Palettes, as they are integral to the generative tools in Canvas.
To import an image from your computer:
Copy and paste
Copy the image to your clipboard (right-click and select Copy, or press Ctrl/Cmd+C).
Paste it onto the canvas using Ctrl/Cmd+V.
Drag and drop
Drag and drop an image from your computer on the Canvas window.
Upload
Click the Upload icon on the left-side panel and select an image from your computer.
Generation windows are bounding boxes used by the Magic Fill, Extend, Remix and Download functions to define the size, aspect ratio, and placement of the image to be generated or downloaded.
Resizing generation windows:
By corners: Preserves the actual aspect ratio.
By sides: Snaps to the nearest supported aspect ratio.
Each generative function produces four images per prompt.
Generate and Remix: Places all four images on the canvas, each in its own image frame.
Magic Fill and Extend: Places one selected image on the canvas, maintaining the size and position of the generation window used.
1. Make sure the Select tool is active in the left-side panel.
2. Click on the desired image frame to display a menu below it.
3. Use the arrows in the menu to scroll through the four generated images.
If the menu doesn’t appear, try refreshing the page.
Image frames can be freely resized and repositioned on the canvas.
The Generate function in Canvas works the same way as it does outside of Canvas. You provide a detailed description of what you want to see, adjust some options, and the AI generates four images based on your input.
The Remix function in Canvas is similar to the standard Remix but offers more flexibility. You can select a specific portion of an image or composition using the generation window and remix just that area.
Suppose you want to create a close-up of a particular image, such as the face of a thinking man.
Click the Select tool on the left-side panel.
Use the generation window to adjust the bounding box around the part of the image you want to remix. You can change the size and aspect ratio.
Enter a prompt that describes the desired result and adjust the options to your liking. Modify the Image Weight slider as needed.
Click Remix. Four images will be generated on the canvas, placed below the existing images.
Magic Fill, also known as inpainting, allows you to modify a part of an image while keeping the rest intact.
Imagine you have an image of a person holding a glass of water, and you want to change it so the person is holding a glass of milk and wearing a watch on her left wrist.
Here’s how:
Click on the Magic Fill button on the left-side panel.
In our case, we used the rectangular tool to mask the hand holding the glass of water and the brush tool to mask the area where the watch should be placed.
Mask the hand holding the glass and the area where the watch should be using the masking tools at the top of the canvas.
Then click Next to get to the next step. This will bring the Magic Fill options at the top of the canvas.
Click Next to proceed.
Adjust the generation window to include both masked areas to fill, and unmasked image for context.
Enter a prompt that includes what appears in the generation window and the changes you want to make. For example:
A photo of a middle-aged woman with short hair holding a glass of milk in a modern, white kitchen with wooden countertops. She is wearing a blue shirt, a necklace and a watch...
Adjust any additional options as desired.
Click Magic Fill to generate the images.
Select the image and use the arrows at the bottom menu to view all four generated options.
Extend, also known as outpainting, lets you expand your image beyond its original edges. It can also modify existing parts of the image, similar to Magic Fill.
Continuing from the Magic Fill example, suppose we want to add a man cutting vegetables on the counter to the left side and change the glass of milk into a cup of coffee.
Click the Extend button on the left-side panel.
Reposition and resize the Extend generation window to your liking.
Mask the hand and glass of milk to indicate they should be changed.
Enter a prompt that describes:
The part overlapping the original image.
What should be generated where the image will be extended.
The changes to be made to the masked areas.
In our case:
A photo of a middle-aged woman with short hair holding a cup of hot coffee in a modern, white kitchen with wooden countertops. She is wearing a blue shirt and a necklace. A man is chopping vegetables on the counter beside her.
Adjust any additional options as desired.
Click Extend to generate the images.
Select the image and use the arrows to view all four generated options and choose the one you like.
To download your final composite image, use the Download function, which works like an adjustable screen capture.
Click on the Download button on the left-side panel.
Alternatively, if you have activated the Always show download outline option, click the download icon at the top left of the inactive download generation window on your canvas.
Resize and position the download generation window to capture the final image you wish to download.
On the left side, near the Download icon, a small panel will appear with the final dimensions of the image. You can adjust the size by manually editing the width and height, then click the arrow next to the Download button to choose the image format (for highest quality PNG downloads, you'll need Ideogram Basic or above. Finally, click Download to save the image.
Click Download to save the image.
Here's the final image.
Consider the following aspects regarding image sizes on the canvas:
Canvas grid: The canvas has an underlying pixel grid, visible when zoomed in at 1600% or more.
Image placement: When uploading or generating images, they will at first align with the canvas grid, with one pixel on the image equaling one pixel on the grid.
Generation window sizes: When you use the Remix, Magic Fill, or Extend tool, the generation window will adopt one of the default predefined aspect ratios, each with a fixed pixel size. The generated image will have this specific size, regardless of how you resize the context frame on the canvas.
Image scaling: Images retain their original width and height in pixels, even if resized on the canvas.
Download dimensions: The Download tool uses the canvas grid to define the size of the image to be downloaded.
The most common aspect ratios and their corresponding sizes can be found in theAspect Ratios & Resolutions section of the documentation.
Let's start with an image generated with this prompt:
A photo of an elderly woman with white hair wearing a floral dress and a necklace. She is seated in a wooden chair near a window in a dimly lit room. The room has a patterned rug, a few pieces of furniture, and some decorative items. The window has a lace curtain. The background reveals a lush green plant and a few objects. The overall ambiance is warm and nostalgic.
Suppose you want to change that person’s face and hands. Using Magic Fill effectively involves understanding image sizes:
Mask the head and hands using the Magic Fill tool.
Resize the generation window to be smaller than the original image but large enough to include something significant to describe in the prompt.
Why make the generation window smaller? Using a smaller generation window provides the AI with more pixels to work with than using the entire original image size. This results in higher pixel density for the selected area, leading to more detailed face and hands.
But keep in mind that some of these details or apparent sharpness may get lost because everything will be interpolated or extrapolated to the pixel grid on the canvas. More on that in this note below.
Adjust the prompt based on the generation window:
A photo of an elderly woman with white hair wearing a floral dress and a necklace. She is seated in a wooden chair in a dimly lit room. The background reveals lush green plants. The overall ambiance is warm and nostalgic.
Click Magic Fill.
Here is the result:
When downloading your composite image, it’s important to understand how scaling and pixel density affect the final output because the final image is based on the canvas’s pixel grid. This means that any images you’ve scaled up or down will not retain their original resolutions in the downloaded image. Enlarged images won’t have extra detail, and reduced images won’t keep their apparent increased sharpness.
Suppose you place a small photo (say, 200 x 200 pixels) on the canvas and enlarge it to double its size (400 x 400 pixels):
On the canvas: It looks bigger, filling more space.
When downloaded: That part of the image may appear blurry or pixelated because enlarging doesn’t add new details; it just makes the existing ones bigger and less clear. The small photo doesn’t have enough detail to look good at a larger size. Essentially, it won’t act like the Upscale function does.
Now, you place a large photo (800 x 800 pixels) and shrink it down to half its size (400 x 400 pixels):
On the canvas: It looks smaller but very clear.
When downloaded: This area will loose its sharpness and some details because the grid will have to transform an image from 800 x 800 pixels to 400 x 400 pixels, loosing details in the process.
In simple terms: The final image you download captures everything within the download area roughly as it appears when the canvas is viewed at 100% zoom, but the overall resolution is determined by the canvas grid.
For best results
Avoid excessive resizing: Don’t enlarge small images too much or shrink large images significantly.
Use appropriate generation window sizes: When generating images, try to use generation windows that are not overly big or small relative to the canvas' pixel grid. Try to reach a good balance when, for example, you are fixing defect part of an image by using a smaller generation window than the original image.
From an image: When viewing an image in detail, click the More icon and select Edit in Canvas.
By default, each new canvas is named “Untitled canvas.” To rename it, click on the canvas name at the top-left corner, next to the Menu icon , and enter your desired name.
Click the Menu icon , to display a list of your recent canvases. Click the icon again to hide the list. From this list, click on any canvas name to access it.
