OpenAI Releases DALL-E 2 Plugin Integration With ChatGPT
OpenAI has recently integrated the DALL-E 2 plugin with ChatGPT, opening new doors for creating images directly within your conversations. The post lays out a simple guide on enabling this feature and sheds light on its potential, especially for content creators. The application and impact of this subtle yet significant ChatGPT enhancement will be seen in the upcoming months, and it's very exciting.
Justin Gluska
Updated June 5, 2023
penguin on the moon
Reading Time: 2 minutes
With the introduction of plugin integration to ChatGPT over the last few weeks, OpenAI has invited a select audience to test out the integration of DALL-E 2 directly within ChatGPT.
You can now directly interact with DALL-E within the chat to create images based on context, ideas, or direct prompting. This is great for those that use ChatGPT to write blogs or prime it with content for video/content production.
We generally use DALL-E to create the featured images in both of our blogs, and now that's even easier because we can paste our entire blog into ChatGPT and ask it to "create 3 featured blog post images based on the context of this article."
It's awesome. I'm awaiting an integration from Midjourney next, but that's a story for another day!
Using DALL-E 2 With ChatGPT
You'll have to turn on plugins if you don't already have them on. Go to your account settings in the beta features tab, and enable plugins.
Once you've done this you can go back to the main screen (GPT-4) and select plugins. Inside here you'll see a plugin store, listing everything on the ChatGPT "app store."
Find the DALL-E plugin, install it, and head back to the chat interface.
You can easily prime it with anything you want and watch the request unfold directly in front of you. You'll see the request, response, and a visualization of the image (that you can do what you wish with)
Very nice, it looks great. It's the same quality and training of images you'd get if you used the website version to create, but the dimensions are automatically wider. By default, it seems like the image size is 1792x1024.
Don't be scared to ask it for something more direct. The default image is more widescreen but you can ask for it in a square. The request actually changes the "image_shape" characteristic of the request. Now you're left with a new image, but cropped.
I will say it didn't take the original image and crop it, but it generated me one that is fairly similar just in square form. Check it out:
So that's it. Need some ideas of how DALL-E could be useful within ChatGPT? We got you covered. Have any questions or had the opportunity to test this out yourself? Leave a comment to let us know how it went!
Want to Learn Even More?
If you enjoyed this article, subscribe to our free newsletter where we share tips & tricks on how to use tech & AI to grow and optimize your business, career, and life.