https://cms.cogniwerk.ai/assets//bc5bda9f-37a7-4c3a-b0d8-b8d58a23bf43.JPG
Super Prompts for ChatGPT: Generate Images and Ideas

Author: 

CogniWerk Editor

Date: 17.10.2023

Want to get Stable Diffusion prompts using ChatGPT? Stable Diffusion is by far the best open source AI image generator on the web. Like most AI tools, it uses prompts, but always brainstorming prompt ideas is not practical. That’s why, in this blog post, we will show you how you can easily generate Stable Diffusion prompts for yourself using ChatGPT in this step-by-step guide. Follow through until the end for some very helpful and easy to copy/use Super Prompt templates!

 

 

What is Stable Diffusion And Why It’s Hard to Create Prompts

Stable Diffusion, a groundbreaking deep learning, text-to-image model developed in 2022, offers a new avenue for generating high-quality images based on text descriptions. However, crafting effective prompts for Stable Diffusion can prove challenging due to the level of specificity and detail required. Each prompt needs to consider various factors, including the subject, medium, style, artist, website, resolution, additional details, color, and lighting. This is where OpenAI’s GPT models, like ChatGPT, can play a significant role. Leveraging the capability of GPT models to generate detailed and creative text, users can automate the process of creating detailed prompts suitable for Stable Diffusion. By feeding the model with a brief description or a set of keywords, users can get back an elaborated and precise prompt, ready to be used in Stable Diffusion. This way, even those with limited experience can easily generate AI image prompts tailored to their needs. Let’s check out how you can create Stable Diffusion prompts with ChatGPT.

 

Step 1: Accessing the ChatGPT Platform

You can start making steady diffusion prompts by going to chat.openai.com, which is the OpenAI website for ChatGPT. To ensure a fresh start, it’s crucial to click the “+ New Chat” button found on the left-hand panel. This initiates a new chat, enabling you to avoid any potential influence from previous conversation history.

Step 2: Crafting a Thorough Prompt Request

Now that you’ve successfully initiated a conversation with ChatGPT, you must articulate your request carefully. Let’s assume you’re seeking to produce prompts for a detailed illustration portrait, follow these sub-steps:

  1. Explicitly Define Image Style: This involves pinpointing the exact style of image you envision. If, for instance, you’re keen on a “photorealistic close-up illustration,” this requirement needs to be clearly stipulated in your request. Such precision helps eliminate ambiguity and enhances the AI’s understanding of your desired output.
  2. Establish a Fixed Subject: A critical aspect of consistent output involves designating a fixed subject. This element can span across a broad range: from objects, animals, landscapes, to individuals. For instance, if you choose a specific person such as the Mona Lisa, this subject serves as a potent keyword. It heavily influences not only the appearance of the portrait but also significantly dictates the style and composition of the prompts generated by ChatGPT.
  3. Priming with Examples: To further streamline the ChatGPT's prompt output, provide two or more examples of the desired prompt. This practice essentially offers a blueprint for ChatGPT, enhancing its comprehension of your expectations. These examples should ideally be in line with your chosen image style and subject.

 

Your assembled ChatGPT prompt may resemble:

Examples of high-quality prompts for a stunning close-up photorealistic illustration of Ana de Armas for text-to-image models (Stable Diffusion, Midjourney, or DALL-E 2) are:

- "A dreamy, vibrant portrait of Mona Lisa; aesthetically pleasing anime style, trending on popular art platforms, minutely detailed, precise, sharp lines, award-winning illustration, 4K resolution, inspired by master artists Eugene de Blaas and Ross Tran, vibrant color palette, intricately detailed"

- "A portrait exuding Alberto Seveso and Geo2099's distinctive styles; an ultra-detailed and hyper-realistic portrayal of Mona Lisa, designed with Nick Knight aesthetics, featuring popular art elements like butterflies and florals, sharp focus, akin to a high-quality studio photograph, meticulous detailing, made famous by artists Tvera, Wlop and Artgerm"

 

For example, when we gave ChatGPT more examples and some super prompts to act as a Stable Diffusion prompt generator, it provided us with many prompts like these:

portrait of Mona Lisa eating pizza, intricate, elegant, glowing lights, highly detailed, digital painting, artstation, concept art, smooth, sharp focus, illustration, art by wlop, mars ravelo, greg rutkowski

 

Step 3: Evaluating ChatGPT’s Output

Following your comprehensive request, ChatGPT will respond with a collection of innovative prompts tailored to match your specified style and subject. You can directly incorporate these prompts into Stable Diffusion, for example via the AUTOMATIC1111’s WebUI, using your preferred settings. To ensure the exclusion of undesirable characteristics in the generated art, it’s recommended to use an appropriate negative prompt, such as: "codeugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, bad anatomy, watermark, signature, cut off, low contrast, underexposed, overexposed, bad art, beginner, amateur, distorted face"

 

Step 4: Adjusting Prompts for Different Scenarios

ChatGPT’s versatility allows you to adjust your prompt request based on the situation. For instance, if you want to generate full-body portraits instead of close-ups, simply include “full-body” in your question and prompt examples.

 

Your revised request might look like:

Examples of high-quality prompts for stunning photorealistic full-body illustration of Ana de Armas for text-to-image models (Stable Diffusion, midjourney, or Dalle2) are:

- "Full-body portrayal of a jubilant Mona Lisa, detailed anime realism, trending on Pixiv, minute detailing, sharp and clean lines, award-winning illustration, 4K resolution, inspired by Eugene de Blaas and Ross Tran's artistry, vibrant color usage, intricately detailed"

- "Full-body depiction of Mona Lisa, Alberto Seveso and Geo2099 style, an intricately detailed and hyper-realistic image in Lisa Frank style, trending on Artstation, butterflies and florals, sharp focus akin to a studio photograph, intricate details, praised by artists Tvera, Wlop, Artgerm"

 

Step 5: Requesting Additional Similar Prompts

Should a particular prompt captivate you, you can solicit more similar examples from ChatGPT.
A simple request can be: "I find the third prompt useful. Could you provide more examples in a similar style?"

This level of customization is key to making sure that the generated diffusion messages really match what you had in mind from the start.

 

Extra Templates: Super Prompts to Easily Copy

Generating Words via ChatGPT

"Please forget all prior prompts. You are an higly skilled linguistic expert in the field of [concept], and your knowledge is highly sought after. I am interested in learning how to create pictures of different types of [concept]. Please give me a list of words describing [concept] in a semantic manner. Remember this prompt until I ask you to forget."

Generating Prompts

"Provide me with examples of high quality prompt for stunning close-up photorealistic [illustration] of [subject] for text-to-image models are

– portrait of beautiful happy young [subject], ethereal, realistic anime, trending on pixiv, detailed, clean lines, sharp lines, crisp lines, award winning illustration, masterpiece, 4k, eugene de blaas and ross tran, vibrant color scheme, intricately detailed

– alberto seveso and geo2099 style, A highly detailed and hyper realistic portrait of a gorgeous [subject], trending on artstation, butterflies, floral, sharp focus, studio photo, intricate details, highly detailed, by Tvera and wlop and artgerm"

 

Conclusion: Ready to Generate With ChatGPT

And that’s it! You now have a good idea of how to use ChatGPT to make prompts for Stable Diffusion that are complex and full of information. With Stable Diffusion, you’re just getting started on an exciting trip. As you move forward, you’ll find a lot of ways to learn more about this advanced AI method and get better at using it. Don't hesitate to experiment with different prompt styles, subjects, and scenarios to see what kind of amazing art you can generate.

Remember that the latest versions of ChatGPT and Stable Diffusion might bring even more capabilities and improvements, so keep an eye on updates from OpenAI and continue exploring the possibilities of AI-generated art.

Now, armed with this knowledge, go forth and create incredible AI-powered images and ideas with Stable Diffusion and ChatGPT! Unleash your creativity and witness the magic of AI artistry like never before.