Firefly is a suite of generative Artificial Intelligence features for creating image content. It’s available as a standalone website at https://firefly.adobe.com. Firefly lets you use everyday language and simple inputs to generate outputs with generative AI.
This guide introduced staff and students to the Adobe Firefly platform, the Adobe applications into which it is integrated and the various GenAI features that it includes. Each feature covered in this guide includes both a video demonstration/tutorial as well as a step-by-step beak down of actions needed to use the feature. For those
Discover Adobe Firefly: Introduction to the magic of Adobe Firefly.
Adobe Firefly is a powerful tool designed to enhance creativity across various applications. Integrated into Adobe Express, Firefly allows users to create social media posts, videos, brochures, and more using simple text prompts to generate images. This tool simplifies tasks like background removal and ensures brand consistency with access to pre-approved logos, colours, and fonts.
Firefly also provides templates and brand elements, enabling the production of high-quality, brand-compliant content. Firefly’s generative AI capabilities such as generative fill, help quickly extend and modify images, saving time and enhancing productivity.
Adobe Firefly produces visually stunning content quickly and efficiently.
How to generate images from a text prompt in Adobe Firefly.
Adobe Firefly can generate an entire image from a text prompt with the ‘Text to Image’ feature. The workflow is effortless.
Video: Link
- Open an internet browser and navigate to https://firefly.adobe.com
- Start by entering a simple description of an image in everyday language (There’s no need to include instructions, like “create” or “make,”) in the Prompt box or scroll down and click on the ‘Text to image’ card, using the letter option allows you to select and use one of the many sample prompts or enter your own in the prompt input box.
- Click Generate. Four images will be generated based on the prompt. If you are unhappy with the results, click on the ‘Generate’ button again to generate four more images (also consider refining your prompt)
- If you choose to refine the prompt by adding additional descriptions, Adobe Firefly displays suggestions to apply to your prompt. It is possible to disable these suggestions (but only once an initial prompt has been executed) by clicking on the ‘Prompt suggestions’ button.
- There is also a series of buttons on the left of the screen where it is possible to further customise the images being generated by Firefly. These include:
- Aspect Ratio of the image
- Content Type (between Photo or Art). By default, the ‘Auto’ option is enabled meaning that Firefly will determine which option it feels is best based on the content of the prompt.
- Hovering over one of the generated images brings up additional options including the Edit button. Clicking on the Edit button provides access to more options including ‘Generate similar’. This will generate three new images based on the selected image.
- For variations of generated images, it is also possible to click on the favourite icon to save this image to your favourites. It is then possible to view favourites on the View tab in the home screen of Firefly.
- To make additional changes to the image, click on the edit ‘Pencil’ icon and take advantage of the generative fill, a composition reference, style reference, add text, add shapes of graphics and create a social media post.
- Once an image has been generated that meets the requirements, click on the download button to download the image in question or select the upscale button to produce an upscaled version of the image(s) that can then be downloaded or edited as previously described.
How to add content to an image in Adobe Firefly.
Using Firefly, it is possible to add anything to an image that can be described in a text prompt using the ‘Generative Fill’ feature. Generating objects, characters, or scenery will save time that would be otherwise spent searching for the right item or creating it by hand. There are also numerous copyright implications that no longer need to be considered or worried about. Use the following workflow.
Video: Link
- Open an internet browser and navigate to https://firefly.adobe.com
- Select the ‘Generative fill’ box or its associated button.
- Upload an image that requires additional content added using the ‘Upload image’ button.
- Using the default selected ‘Insert’ option, select where you want new content added by brushing over part of your image to let Firefly know where to add content.
- Enter a text prompt that describes what you want Firefly to add.
- Click the ‘Generate’ button.
- Firefly will generate three new items that blend perfectly with your surrounding image. Choose the variation you like best and click Keep. If none of the generated images are suitable, either press the ‘More’ button or select the Cancel button and refine the prompt.
How to replace an image background in Adobe Firefly.
It is possible to generate a new background with the ‘Generative Fill’ feature available in Adobe Firefly.
Video: Link
- Open an internet browser and navigate to https://firefly.adobe.com
- Select the ‘Generative fill’ box or its associated button.
- Upload an image that requires the background to be removed or replaced using the ‘Upload image’ button.
- Using the default selected ‘Insert’ option, click the ‘Select background’ option, this should remove the current background, leaving only the aspects of the image that are considered in focus in the foreground.
- If any of the original background is not automatically removed, click on the ‘Brush settings’, reduce the ‘Brush size’ and increase the ‘Brush hardness’ to make the brush edge sharper.
- In the prompt, describe the background that must be added to the image and click ‘Generate.’
- Firefly will generate three possible background options. You can select one of the backgrounds, regenerate to produce three new versions or refine your prompt.
How to remove content from an image in Adobe Firefly.
Just like the ability to introduce new content to an image, it is also possible to remove specific content and have the space this leaves match the original image. For example, removing a tree from a field would replace the area the tree filled with a consistent-looking area of land to the rest of the field.
Video: Link
- Open an internet browser and navigate to https://firefly.adobe.com
- Select the ‘Generative fill’ box or its associated button.
- Upload an image that requires content removal, using the ‘Upload image’ button.
- Select the ‘Remove’ option located in the left-hand menu and select the content to be removed by brushing over part of your image.
- Click the ‘Remove’ button.
- Firefly will generate three new versions of the image with the selected content removed that blends with the original image.
- Choose the variation you like best and click Keep. If none of the generated images are suitable, either press the ‘More’ button or select the ‘Cancel’ button and refine the prompt.
How to extend an image in Adobe Firefly.
It is possible to enlarge the canvas of an image and add missing details or entirely new elements in the expanded area. Remember that Adobe Firefly’s GenAI ‘Expand’ feature can only assume what details should be in the expanded region of images and its output will in almost all cases, not be truly accurate.
Video: Link
- Open an internet browser and navigate to https://firefly.adobe.com
- Select the ‘Generative fill’ box or its associated button.
- Upload an image that requires content expanding, using the ‘Upload image’ button.
- Select the ‘Expand’ option located in the left-hand menu.
- Select from the available ratio pre-sets or use the handles that appear in the corners, top, bottom and sides of the image to drag out the space into which the image is to be expanded.
- Add a description to the Prompt field of what content to be included in the extended background or leave it blank if you want it to be extended based on the existing image.
- Click the ‘Generate’ button.
- Firefly will generate three new versions of the expanded image that blend with the original image.
- Choose the variation you like best and click Keep. If none of the generated images are suitable, either press the ‘More’ button or select the ‘Cancel’ button and redefine the area into which the image should be expanded.
How to transform text into creative images.
Enhance your creative process by integrating structure and style references to accurately tailor your results. This workflow is particularly useful as GenAI image generation can still struggle with generating accurate spelling of words.
Video: Link
- Develop a prompt for the creative image (that includes text) you want to generate. “Ice block lettering in the snow at night with the northern lights in the background”.
- Either upload or generate an image that includes the text or logo as a Structure reference This will guide the image generation process, whilst still maintaining creative freedom.
- Use the Strength slider to control how closely the result resembles the reference image used. Drag the Strength slider to the right for stronger influence; drag the slider further left for more creative variation.
- In the Styles section, upload a reference image with the desired style (photo, painting, sketch, etc.) or browse the Gallery to control the overall visual appearance. Firefly will apply a similar style to your generated visuals.
- Use the Strength slider to control how closely the result resembles the style image used. Drag the Strength slider to the right for stronger influence; drag the slider further left for more creative variation.
- Once the required prompt, structure reference and style reference have been added, click the generate button.
For further support and guidance, please contact the Digital Learning team.