{
"subject": "an adult woman with an original face, bearing no resemblance to any real person, fair skin, dark brown hair tied up in a high bun with loose strands on her forehead, thick eyebrows raised high, eyes wide open and mouth open in a delighted shock that shows her upper teeth; she wears a plain wine coloured crew neck t-shirt; she faces the camera, cropped just below the chest, with her left hand open and raised beside her face, fingers spread and palm turned towards the viewer",
"cutout": "the woman is cut out from the background like a sticker pasted on top of the artwork: a thin white outline of even weight follows her whole body, arm, hand and hair, and behind that outline sits a black shadow offset to the right and downwards",
"product": "in the right third, a deep white ceramic bowl, cropped by the bottom and right edges of the frame, full of spaghetti coiled in bright red tomato sauce, with grated cheese and green leaves scattered on top and three curling wisps of white steam rising; the food looks larger than life, very glossy and very saturated",
"typography": "the ONLY words written in the image are I TRIED and THIS, one under the other, each in its own black rounded box; heavy condensed sans serif capitals, tilted slightly to the right, in strong yellow with a thick black outline and a hard offset black shadow; the lower word is wider than the upper one and reaches a little over the bowl",
"scene": "a flat background with no recognisable place: a radial gradient running from orange red behind the woman's head to almost black deep red at the edges, with lighter diagonal rays coming out of the centre and a dark vignette closing all four corners",
"composition": "a horizontal image in 16:9; the woman fills the left third and the bowl the right third, both cropped by the bottom edge; the block of two words sits in the middle third, centred vertically; the bottom right corner stays free of text and of any element",
"camera": "straight on framing at eye level, no lens distortion, everything in focus from front to back",
"lighting": "strong frontal studio light on the face, with no hard shadow on the skin; a bluish rim light on the shoulders and the top of the hair separates the figure from the warm background; the food takes a warm side light that picks out the shine of the sauce",
"color_palette": "red and orange in the background, yellow in the text, wine in the t-shirt, white in the bowl and in the cutout outline, green only in the leaves; high saturation and high contrast",
"style": "a YouTube video thumbnail: a treated photo pasted over a drawn background, clean and eye catching, made to be read on a phone screen",
"details": "the closed white outline around the woman, the hard shadow behind the text boxes, the wisps of steam off the food and the empty bottom right corner",
"negative_prompt": "any word other than I TRIED and THIS, a stray letter, a number, a date, an acronym, a watermark, a logo, a real brand, a likeness of a famous person, a child's face, small text, a caption, a frame, a decorated border, washed out colours, soft focus, a hand with an extra finger, a deformed hand, AI artefacts"
}
How to use
A video thumbnail is what decides whether anyone clicks: the title comes second. This prompt builds the art in the wide YouTube shape, with a big face, three words you can read on a phone and the right hand side free for the object of your video. It works in ChatGPT, Gemini and NanoBanana.
1. Swap the words in the typography field for your own. Three at most, in capitals, and say they are the only ones in the image.
2. Change the subject field to the face and the expression that fit your video. A neutral expression is what costs the most clicks.
3. Check the letters before you use it. The generator swaps letters in long words: short words go wrong less often.
4. Look at the art shrunk to the size of a fingernail. If the text disappears there, it disappears on the phone of anyone scrolling.
5. Make three versions changing only the color_palette field. The same art in another colour moves the click rate more than swapping the photo.
Tip: leave the bottom right corner empty. That is where YouTube itself stamps the length of the video.
🍪 We use cookies. Some are essential for the site to work. Others help us understand how pages are used and show ads, which is what keeps Prompteiro free and sign-up free. Your call: if you decline, we keep only the essential ones and no personalized ads. Read the Privacy Policy.