To use these prompts for Gemini AI Halloween photos or in ChatGPT, attach your photo first and paste one complete prompt, in four steps.
-
Pick a look. Choose a look above, such as Moonlit Unicorn, and press Copy prompt to copy all five blocks at once.
-
Attach your photo first. Open an image-editing chat in Gemini or ChatGPT and attach a clear selfie or pet photo before you paste anything.
-
Paste the full prompt. Paste the copied prompt as one message, keeping PRESERVE and EXCLUDE in place.
-
Change one line at a time. To make the look yours, edit one colour in CHANGE, such as "moss-green" to "rust-orange", and generate again.
These prompts are written for any image editor that accepts a reference photo, including Gemini and ChatGPT. Without a photo, the editor has no face or pet to keep, and the PRESERVE block has nothing to work on. Features and limits depend on the service and your account, and photos you upload there follow that service's data policy.
Gemini and ChatGPT set the picture shape and take your photo in different ways. This is what their own help pages say:
Sources: Google's Nano Banana Pro prompting tips, Gemini Apps upload help and Gemini Apps image help; OpenAI's Images in ChatGPT and uploading images.
The same prompt can give different results in different editors, and the example images were made with Shelfcanvas. If an editor refuses a prompt, try removing any words you added, rather than deleting the EXCLUDE block. Not affiliated with Google or OpenAI.