We're excited to announce GPT-2 Image, a new way to generate images with GPT-2! 🤖

OpenAI's next-generation image model with state-of-the-art photorealism, pixel-perfect text rendering, and brand-consistent product photography.

Available in Prompter and Mixer2.

We're excited to introduce Mixer 2, a new take on image mixing! ✨

Upload up to 4 images, add some text guidance, and Mixer 2 will generate creative concepts — each with its own AI-crafted prompt and images. Think of it as a collaborator that takes your references and brainstorms where to take them.

You can vary concepts to explore new directions, edit prompts, and choose from multiple generation models, including the latest like GPT Image and Nano Banana 2. Everything is saved in your session so you can keep iterating.

We've added some of the latest and greatest models to Prompter for image editing and generation! You'll find them in "Instruct" mode.

Flux 2 and Flux 2 Flash

Nano Banana Pro

GPT Image 1.5

We're excited to announce video generation with sound! 🔊

Our latest model for Animator, Wan 2.5, brings enhanced video quality, smoother motion, and improved prompt responsiveness. And native audio too!

We've added a few new models to help you refine and re-imagine your images. These models are designed to enhance your images, allowing you to fix problems, adjust details, add elements, or even change the style of your images.

In-Paint: Select an area you want to edit and provide your prompt.

Instruct: Describe the changes you want to make.

We know the NSFW system can be pretty frustrating when it gets it wrong. We've just added the ability to un-flag your own images 🚩

This calls a more powerful and costlier checker so we can't do it for every image. To prevent abuse, free users get 3 free unflags a day and subscribers get unlimited.

You can unflag your own images by clicking the flag icon on the image and then clicking "Unflag" in the dropdown.

We're excited to announce our newest tool: Animator 📽️

Transform your static images into captivating videos by combining them with text prompts. This powerful feature adds a new dimension to your artwork, bringing your creations to life with dynamic motion and effects.

We've selected high-quality models while keeping costs manageable, making Animator available exclusively to our subscribers. Start exploring the possibilities of motion and storytelling with your images today!

We're excited to announce Splicer 2, our new tree-based interface for mixing, editing, and exploring images! 🌳🧬

This powerful tool lets you create and navigate through branching image variations, making it easier than ever to explore different creative possibilities. Plus, we've added an achievements system to reward you for discovering unique and interesting images.

Introducing Paint Mode inside Composer 🎨

You can paint and image to regions of the canvas to control exactly where it should go. These can be overlapped to create interesting mixes and compositions. This is very useful for getting two different faces in one image.

We've made a few updates to Collager:

1. You can now add Artbreeder images from within Collager

2. Drawing now has a new "paint brush" mode, along with the original "shape" drawing mode.

(These updates will work in Composer+Collage as well)

Now you can create and use collages from within Composer. This allows the best of both words - spatial control of Collager but also the style/face control of Composer. ✂️ 🎚️

To use it go to Composer ➡️ Add ➡️ Composition ➡️ Collager

We've added a user directory page for searching and discovering other users! The Most Active Users of the day are computed by credits spent per day. Users who are mostly working in private mode are not shown.

We're launching a new tool called 'Composer,' which is our most controllable and fastest tool yet. You control which images should be used for style, content, or composition, and images take around one second to generate. We are also making image generation free for subscribers of any tier. Images can be enhanced from within the tool, which costs 0.1 credits for all users.

There are more improvements coming soon that we look forward to sharing. This includes deprecating some tools such as Mixer, Tuner, Poser, and Prompter and merging them into a single, unified, and easier-to-learn tool.

We've launched a new feature called "Remix Party." It's a social space to collaborate by remixing images together. 🔀 🎉

Every day we start with a different image as a seed of inspiration. As people share their remixes, everything that's shared is fair game for remixing. We remix the remix's remix... and so on.

From this single image, the remixes will branch out and take us on a strange and wonderful journey to who knows where.

The most recent Remix Party was joined by 60+ artists over 24 hours, and we made 300+ remixes. This image shows just a few of the branches the remixes took. We hope you'll join us for the next one!