You can now create a unique New Year postcard with your own face without resorting to professional designers. With the help of a special prompt (instruction) for neural networks, ordinary photos can easily be transformed into atmospheric cards inspired by Soviet fairy tales of the 1960s–1970s.
To achieve high-quality results, it is recommended to use advanced language models such as GPT-5.2 Thinking. This neural network is capable of building complex chains of reasoning, allowing it to preserve facial features, proportions, and natural skin texture as accurately as possible. The result is not just a stylized image, but a recognizable portrait in festive retro aesthetics.
Ready-made prompt for the neural network
To get the result, copy the text below, upload your photos (preferably 1–3 clear images with good lighting), and send them to a chatbot with image generation capabilities:
Prompt text:
“You are an art director creating vintage Soviet postcards. Determine the number of people in the photo and depict EXACTLY the same number, preserving their real appearance as much as possible without retouching. Style: vertical New Year postcard from the 1960s, fairy-tale atmosphere, soft color palette. Details: winter clothing (fur hats), fir branches, pine cones, realistic forest animals. Text: clear inscription ‘Happy New Year’. Requirements: no modern objects or logos, no extra faces or distortions.”
This method helps solve the problem of finding original holiday greetings. Instead of using standard images from the internet, you create a personalized digital gift in which your loved ones become the heroes of a warm winter story.
If the first generation is not perfect, you can give the neural network additional instructions in a dialog mode—for example, ask to change the clothing or add more snow. Modern models adjust the image while taking feedback into account and preserving the overall composition.
ORIENT
