Quickstart guide & FAQ
Seven steps from blank page to model-ready prompt. Takes about two minutes once you've done it once.
Choose the image model you'll run the prompt in — ChatGPT, Gemini, Nano Banana, Meta AI, or Universal. The prompt is worded for the model you pick.
Tell it what the new image should depict — A lighthouse on a cliff at dusk. Leave it blank and the model invents a completely unrelated subject and world.
Transformation distance, 1 to 5. Level 1 keeps the source recognizable; level 5 extracts the DNA so far the source vanishes. Level 4 — DNA Only — is the default sweet spot.
Strict means only your selected DNA survives — everything unselected gets deliberately discarded. Turn it off for a looser, more flexible interpretation.
Photorealistic, cinematic, painterly, surreal — toggle the pills to set how the finished image should render. The defaults give you photorealistic, high-resolution, camera-real output.
Expand the categories — Light, Color, Texture, Composition, and nine more — and check the visual relationships you want extracted. These are relationships, not objects: light direction, not lamp. New here? Hit ✦ Surprise Me first for a fresh DNA selection — see what it can do, then tweak any choice.
Hit Copy Prompt. Then open your chosen image model, attach your source image, paste the prompt, and generate. Your image never comes here — it stays with you the whole time.
No. Your image never comes to this site. You look at your own image, decide which visual relationships matter, and the tool builds the prompt. You then use that prompt with your image inside whatever image model you chose.
Refresh the page. The tool runs entirely in your browser, so a refresh reloads it clean. One heads-up: refreshing resets your selections, so copy any prompt you like before you do.
How far the new image travels from the source. At level 1 (Reimagine), the source may still be noticeable. At level 4 (DNA Only), the source should be unrecognizable — only the selected relationships survive. At level 5 (Extreme Extraction), even the relationships get pushed toward abstraction.
Strict is the default and the honest one: only selected DNA may survive, and everything unselected is explicitly named as discarded. Flexible allows minor incidental source influence when it doesn't make the source recognizable. When in doubt, stay strict.
Midjourney wants a compact, single-block prompt format rather than the sectioned layout the other models get. That variant is built and will return to the dropdown later.
No account, no cost, no sign-up. The tool builds prompts in your browser session. You'll need your own access to whichever image model you target — that's between you and them.
Creative DNA is a BRENTCORP studio tool — built in the open, one tool at a time. Read the story at BRENTCORP HQ.