Use ChatGPT to Create Images for Your Blog Posts (DALL-E 2 Integration)

With the release of plugins in ChatGPT, you can now integrate DALL-E 2 directly with ChatGPT, allowing you to create images for your blog posts extremely easy.

Justin Gluska

Updated June 4, 2023

A DALL-E-generated image based on a prompt and 3 variations of it

A DALL-E-generated image based on a prompt and 3 variations of it

Reading Time: 5 minutes

In the world of blogging, a picture is worth a thousand words. The right image can capture the essence of your blog post, engage your readers, and even boost your SEO. But finding the perfect image can not just cost money, it can also be a time-consuming task.

Personally I'm pretty lazy when it comes to featured images. It's not that they aren't important, but they can be automated or sped up (at least for what I write about). If you run a food blog, you should probably use your own food pictures. But we write about robots. I'm not sure where to find one of those in real life.

ChatGPT seems to be getting updates every 2-3 weeks at this point. It's great. They just released plugins for everyone – letting you add third-party apps to really expand the potential of the chatbot.

I've been using tools like DALL-E 2 and Midjourney to create stock images for my blogs and various publications I need images in. They're great. Both work better for different purposes but still aim to achieve the same goal – creating images from a line of text.

If you write blogs or produce content for anyone and need original, high-quality, and customizable featured or post images, here's how you can reduce your article production time by at least 5-10 minutes a post using ChatGPT.

Producing DALL-E 2 Images with ChatGPT

So the first thing you'll need is the premium version of ChatGPT (the $20 month plan). It's the only way you'll get access to all of the third-party plugins that OpenAI enabled fairly recently.

Once that's done you'll need to click the dropdown under GPT-4 (BTW: you can't use plugins with GPT-3.5 and probably won't ever be able to)

Go ahead and turn plugins on.

GPT-4 plugin beta feature on the dashboard of ChatGPT

If you don't see this option in the dropdown you'll have to enable it in the settings (click your email on the bottom left and select settings). Then go into beta features and turn plugins on.

Enable plugins within the ChatGPT settings and beta feature element on the ChatGPT homepage.

Ok now that you've gotten that out of the way. Let's install the DALL-E plugin.

You might see a little icon when you enable plugins. Click that and open the plugin store. Install DALL-E. And I promise that's it!

A list of plugins in the ChatGPT plugin store showing DALL-E as one of them.

So how do you use this when writing blogs? Well once the plugin is installed you can literally just ask. It seems to produce images with a 3:2 ratio, breaking the standard square box boundary used within the default DALL-E.

Paste Your Blog and Ask for An Image

So I'll take an existing blog on natural language processing, for example. Go ahead and copy a majority of the article into ChatGPT and ask it to create 3 featured blog post images based on the context of the article you've pasted.

A sample blog post with text that should be copied to put into ChatGPT
A conversation with ChatGPT being given context of an article

Now sit back, let ChatGPT make a request to generate some pictures, and take a look at your new featured images. You can actually expand the request and see the exact prompts that are being used to create your pictures.

The DALL-E 2 request within ChatGPT showing the image prompts used to generate an image

Once this is done you'll get a JSON code-looking response with images, but if you just scroll down you'll see your images embedded into ChatGPT. You could right click these and save them, then publish them on your blog. Based on a description I gave, these are the images I got back:

An image generated by DALL-E 2 from the context of a blog
An image generated by DALL-E 2 from the context of a blog
An image generated by DALL-E 2 from the context of a blog

I really like the second one. If I were to replace the current featured image in the existing article, I'd use that.

Now imagine how much faster you can post blogs. I'm not sure if this actually saves a ton of time, but it really really helps me out.

Ask for Any Type of Image Without Context

If you aren't writing a blog (or just don't want to prime ChatGPT with the content of it) you can just ask for an image. You could be a bit more creative with it if you don't really know what you're looking for. Try to give it something abstract.

Check out the request, and look at these insanely awesome pictures that were produced!

The request from DALL-E within ChatGPT showing 3 image prompts, based on a simple description a user gave to ChatGPT

Picture 1: A serene scene of a golden retriever and a tabby cat sitting side by side on a lush green lawn, bathed in the warm glow of a setting sun. The dog's tail is wagging gently, and the cat is purring contentedly, their eyes closed in peaceful bliss.

DALL-E 2 image produced by a sentence of context given to ChatGPT

Picture 2: A heartwarming image of a husky and a Siamese cat sharing a large cushion in a cozy living room. The fire is crackling in the fireplace, casting a soft, warm light on their fur. They are both asleep, their bodies touching, radiating a sense of peace and coexistence.

DALL-E 2 image produced by a sentence of context given to ChatGPT

Picture 3: An adorable picture of a beagle puppy and a fluffy white kitten playing with a red ball in a sunny backyard. They are taking turns, showing respect for each other's turns, embodying the spirit of peaceful coexistence.

DALL-E 2 image produced by a sentence of context given to ChatGPT

Give a Description & Pick The Best One

Lastly you can actually give ChatGPT your exact prompt and ask it to generate a few variations of a that singlular prompt.

Something like this:

ChatGPT being prompted with creating 3 images based on a prompt
A request to DALL-E to generate 3 variations of a singular image based on one prompt

So you'll actually see ChatGPT changes the 3 image requests that it gets from DALL-E. I believe it takes the context of what you're looking for and adds a few variations within that. You'll still get something very similar to your initial prompt, just asked in a different way. Here's what we got:

A DALL-E-generated image based on a prompt and 3 variations of it
A DALL-E-generated image based on a prompt and 3 variations of it
A DALL-E-generated image based on a prompt and 3 variations of it

And that's pretty much it! This should at least save you a bit of time when writing up blog posts. It's kind of nice to have it all done in the background. If you have an assistant find stock images for you, this is even a better way of having them do so. Hopefully in the future we'll see a ChatGPT plugin for image generators like Midjourney!

What are your thoughts on the current plugins for ChatGPT? Have you used any others? Let us know!

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.


Written by Justin Gluska

Justin is the founder of Gold Penguin, a business technology blog that helps people start, grow, and scale their business using AI. The world is changing and he believes it's best to make use of the new technology that is starting to change the world. If it can help you make more money or save you time, he'll write about it!

Subscribe
Notify of
guest

1 Comment
Most Voted
Newest Oldest
Inline Feedbacks
View all comments