Image generation | Page 2 | CyberSEO Pro | Support Forum

CyberSEO Pro is a powerful tool designed to bring a webmaster's idea to life, not to replace the webmaster's strategic thinking. You're the master, the one in charge of conceptualizing the end product; the plugin simply serves as a tool to execute your plans.

For image generation relevant to news articles, the algorithm or logic must come from you. Can you explain how you envision this process? For example, if you have a random article, what's your algorithm for determining what the image should represent? An AI image generation tool can generate virtually anything, but it needs specific instructions. So how do you, as the webmaster, decide what those specifics should be?

You can use your post title in Stable Diffusion and DELL-E prompts using the %post_title% shortcode. But it won't work, because Statble Diffusion and DELL-E expect an image description, not a post title. Course if your article title says: "Epic sunshine over New York City, high detail, deep perspective, cinematic", it will definitely work. But you won't get anything good for "The government has set a new excise duty on alcoholic beverages with a strength of up to 12%", using %post_title% as a Stable Diffusion prompt.

Try this code. It will ask ChatGPT to read the article and name the object for the image that best describes the content of the text. The script will then pass the ChatGPT response to Stable Diffusion as an image job. The rest of the parameters are up to your taste (find more information Login to see this link). The result will be saved as a "thumb" custom field that you can use to generate a post thumbnail.

Login to see the code

If you want to improve the prompt to ChatGPT – just do it. If you want to use some other Stable Diffusion preset instead of “cinematic”, feel free to set your own.

Since this code might be useful for other CyberSEO Pro users, I am posting it here on the public forum.