CREATIVE GENERATIVE AI
Explore how creative dreams and experiments can become AI creations.
If you’ve ever wondered what a giant octopus invading New York City might look like, generative AI can show you.
There are different kinds of Generative Ai models. Large language models or an LLM, for instance, are designed to create complex texts from just a few words. Firefly is a diffusion model, which is designed to create images from text prompts. The words you use in your prompt act as instructions to the generative AI model explaining what type of content to generate — like an image of a giant octopus on the loose in the Big Apple.
Adobe Firefly is a family of creative generative AI models planned to appear in Adobe Creative Cloud products including Adobe Express, Photoshop and Illustrator. Firefly’s first model is trained on a dataset of Adobe stock, openly licensed content and content in the public domain where the copyright has expired and uses simple text descriptions to generate amazing images, unique text effects and more.
Whether you want an image of an antiquated car engine or a retro futuristic boom box, the text to image feature can generate it for you. And it’s only the start. This Firefly model lets users adjust aspects of the initial outputs with a series of in-app tools.
Many AI-powered image generators offer a single function, but Firefly offers a range of capabilities including text effects and the ability to add and remove elements from an image. In the future, we plan to offer even more features — including tools powered by Firefly within Creative Cloud apps — such as the ability to adjust the mood of a video, remove distractions from photos and adding texture to 3D objects.
Outside of Stable Diffusion, it is important to know the pros and cons of other generative AI tools such as DALL·E and Midjourney when compared against Firefly.
Text-to-image prompt: Retro futuristic boom box with sunglasses.
Adobe Firefly trains on Adobe Stock imagery in accordance with the Stock Contributor Licence agreement, openly licensed content and public domain content with an expired copyright. This means that Firefly is designed to be safe for commercial use and Firefly outputs are now available to be used for commercial purposes.
The Firefly generative AI models are being built to help creators and include them in the process. You can read about our compensation model for Adobe Stock contributors whose content is used in the dataset to train Firefly models. The goal is to create a generative AI model that benefits creatives and responsibly speeds up creative work.
Turning ideas into visuals takes time for even the fastest creators. Adobe Firefly can help veteran artists, rising stars, imaginative writers, enterprise teams and savvy small business owners alike when it comes to creating and communicating concepts.
For example, imagine you’re a website designer working for a recycled clothing company. You head to Firefly and enter the prompt “baseball hat made of recycled denim on a leafy green background” to create a mood board to show your client what you envision. They love the idea, but also want to see an option with a yellow flannel. Adjusting your text prompt to “shirt made of recycled yellow and black checked flannel hanging from a tree branch” gives you another proof-of-concept to ensure your photographer goes on on site with a shot list to capture exactly what the project requires.
Fast AI art, image and photorealistic examples can keep you from wasting time and help your whole team stay on the same page.
Text-to-image prompt: A wormhole opens in outer space and a huge 17th century sailboat emerges.
Once you have your generated creations, whether they are artistic images or photorealistic, you can refine them with tools in Firefly. Choose from styles like cubism and maximalism in the Styles menu, which features different art movements and themes. You can even choose different materials like fur or clay to adjust the texture of the image.
Refine the Lighting options, as well as the Colour and Tone of an image, to explore different looks for your results. And choose from photo compositions like Close up, Shot from above and others to try out a variety of views.
Mix and match options from these menus or stack them together, to create even more possibilities for each text to image prompt.
Text-to-image prompt: Hot air balloon full of flowers and butterflies, orange sky in background, happy birthday basket balloons, cubism, wide angle, golden hour, vibrant colour.
After Firefly’s AI image generator, try adding an image to text with the Text effects model. Choose a word or phrase, like your brand name or social media handle and type it in. Then describe the effect you want to see on the letters or choose from options like flowers, lava, gingerbread and many more. Click the Generate button and watch your words become artistic AI creations. You can also adjust the font, font colour and background colour to try out a variety of creative choices.
To best understand the creative possibilities of Firefly, you should give it a shot. Get inspired by anything and everything you can imagine. Explore what Firefly can help you to create.
Use generative AI and simple prompts to create the highest-quality output - images, text effects, colour palettes and more.
Based on your location, we think you may prefer the United States website, where you'll get regional content, offerings, and pricing.
