How ChatGPT Sees Me, Drawn From What You've Already Told It

How ChatGPT Sees Me, Drawn From What You’ve Already Told It

Be the first to rate 11 🖼️ ImagePortrait

Prompt

{ "memory": "create an image of how you see me, based on what you know about me from memory and from our conversations. Everything that appears in the scene has to come from something I actually told you: nothing made up and nothing generic to fill the frame. If you have no memory of me, draw the example described in this JSON", "before_the_image": "before generating, write out as text the list of the six facts about me that will go into the scene, each one with the object that will represent it. Only what I told you about work, hobbies, tastes, routine and pets goes in; leave out everything about health, money, address, family, relationships, names of people and anything I wouldn't post. Then generate the image with exactly those six objects", "subject": "me, the way you picture me; if I upload a photo, the face, hair, skin and glasses are the ones in the photo. In the example, a man of about 30, with Black skin, short coily hair with volume on top, a thin, neatly trimmed beard, round glasses with thin gold frames, a slight closed-mouth smile, looking down at the paper he is grading", "objects": "the six facts become six concrete objects, each one easy to recognize on its own. In the example: (1) math teacher: stacks of tests on the counter, graded in red pen; (2) sourdough bread: a round loaf with a golden crust and a score on top, on a wooden board, and a glass jar of sourdough starter with an orange rubber band marking the level; (3) running training: a pair of gray running shoes with orange soles and plain sides, no logo, hanging by the laces from a hook near the door; (4) the cat: an orange tabby cat sleeping curled up on top of the tests; (5) learning guitar: a light-wood acoustic guitar leaning in the corner, next to the counter; (6) succulents: a row of small succulent pots on the windowsill", "wardrobe": "everyday clothes for staying home; in the example, a plain dark moss-green short-sleeved T-shirt, with no print", "action": "me sitting at the kitchen counter, grading a test with the red pen in my right hand (on the viewer's left) and my left hand resting on the back of the cat, which is asleep on top of the papers", "scene": "the kitchen of a small apartment in the early evening. Behind me, a large window fills the upper left, with the city outside: tall buildings with lit windows against the dark blue sky of dusk; on the windowsill, the row of succulents and a trailing pothos spilling over the left edge. A wide-shaded pendant lamp at the upper left, switched on. On the right, the light-colored wall with the shoe hook up high and the wooden door on the right edge, with the doorknob visible; the guitar leans just below, to the right of the counter. In the foreground, the dark, worn wooden counter runs across the frame: on the left, the starter jar at the edge and the bread on the board; in the center and on the right, the stacks of tests and the cat. In the bottom left corner, a blue and white checked dish towel; in the bottom right corner, the round seat of a wooden stool", "composition": "vertical in a 3:4 ratio, an interior scene seen from the front and slightly from above, with the person in the center: the head between ~41% and ~65% of the width, the top of the hair at ~15% of the height, the eyes at ~31% and the chin at ~37%; the counter starts at ~58% of the height and runs down to the bottom edge; the cat between ~42% and ~89% of the width, at a height of ~47% to ~62%; the bread and the jar in the lower band on the left, between ~50% and ~77% of the height; the shoes at the upper right, down to ~23% of the height. All six objects sit fully inside the frame, none of them cut off by the edge", "camera": "the point of view of someone on the other side of the counter, slightly above the eye level of a seated person, a waist-up framing with the room around it, everything in focus", "lighting": "warm golden light from the pendant lamp falling on the counter, the face and the cat, against the cold dark blue of the window; warm reflections on the wood and on the glass of the jar, soft shadows under the objects", "color_palette": "gold, amber and orange from the light, the wood, the bread and the cat against the navy blue of the sky; green from the plants and the T-shirt, white from the tests with the red of the pen as an accent", "style": "a cozy painted digital illustration, with visible brushstrokes and paint texture, soft outlines without black lines, medium detail, the look of a magazine illustration; not a photo, not a 3D render and not anime", "details": "three proofs that the image is about ME: each of the six objects points to a fact I told you, and no object is there just for decoration; the person is me (or the man in the example) and not a generic character; and nothing in the scene exposes anything private. The tests have only scribbles and illegible red marks", "negative_prompt": "robot, AI mascot, ChatGPT logo, laptop, coffee cup, object I never mentioned, generic character, someone else's face, written name, legible text, letters, numbers, document with personal data, real logos or brands, anime style, 3D render, realistic photo, objects cut off by the edge, deformed hands, extra fingers, resemblance to a real person, watermark, AI artifacts" }

How to use

The trend of asking ChatGPT for an image of how it sees you tends to come back with the same scene for everyone: a laptop, a coffee cup, a plant and sometimes a little robot, because whatever it doesn't know it fills in with clichés. This prompt makes every object in the scene come from something you actually told it, asks for the list of facts before the image and leaves out anything private. Works in ChatGPT with memory turned on. 1. Check in ChatGPT's settings that memory is on. 2. Paste the prompt into a regular chat: a temporary chat doesn't generate images. 3. Read the list of facts that comes before the image and cut anything wrong or too personal. 4. If you want your own face in the scene, upload a photo with it. 5. If an object shows up that you never mentioned, ask again and say which one. Tip: compare yours with a friend's image; the fun of the trend is recognizing the person from the objects alone.

Works with ChatGPT

Rate this prompt

📖 From the Prompteiro blog

See all posts →

💡 You may also like

Professional LinkedIn Headshot with AI
🖼️ Image★ 5.0

Professional LinkedIn Headshot with AI

Mother and Daughter Photoshoot in Matching Outfits
🖼️ Image★ 5.0

Mother and Daughter Photoshoot in Matching Outfits