Turn your images into custom DreamBooth AI models

Unlike traditional model training or fine-tuning, the DreamBooth process used in our Model Trainer adds a special "instance prompt" token to the base image generation model—in this case, the powerful FLUX.1 [dev]. This token is unique to your custom model and automatically invokes your Character or Style when generating images with the AI Generator.

For instance, you could upload a series of photos of yourself taken from various angles and settings. The AI will then learn your unique features and will be able to generate realistic AI images of you in various scenarios—like at the beach, in a cartoon style, or wearing different accessories.

By leveraging FLUX.1 [dev], the Model Trainer delivers exceptional results, combining DreamBooth’s personalization with the speed, versatility, and precision of one of the most advanced AI models available. This makes it easier than ever to create lifelike characters, niche styles, or entirely unique visuals tailored to your needs.

Once trained, your DreamBooth models can be accessed through the Image Generator. We store them on our servers, and you can use them anytime, anywhere, provided you have an Internet connection. This way, there are no complicated setups or technical requirements for your PC, and you can just jump right into image generation.

To be more specific, custom DreamBooth models are available in AI Generator’s Essential mode, with FLUX.1 [dev] selected as the base model. Select a Style from the menu within the prompt box and/or a Character from the dropdown list in the Consistent Characters section, and start creating. Take advantage of our intuitive interface and lightning-fast generation to quickly experiment with different ideas.

We offer a fast and effective training process on our end, but as with all AI models, DreamBooth needs high-quality input to deliver the best output. The quantity and quality of the training images you upload will significantly impact the model’s performance.

While you technically can upload as few as 3 images, we recommend using between 10 and 20 training photos. Follow our detailed instructions set out in the guide to using Model Trainer to achieve the best results.

Don’t worry; you don’t need technical skills to successfully create a high-performing AI model. There are no technical parameters to set; all you need to do is name the model, select the gender if you’re training a Character, and supply the images.