TechieSA shares a GPT Image 2 prompt for miniature cities in knitted yarn style

On June 19, TechieSA (@TechieBySA) posted to X the series Knitted Cities —woven cities— generated with GPT Image 2 in SocialSight. The thread includes four almost identical prompts replacing the city—Paris, NYC, Venice, and Tokyo—and a answer with the reusable template using [CITY].

AI Prompt · How to do it

Second panel of the yarn: New York in yarn-art. @TechieBySA

Steps

1. Open SocialSight or ChatGPT with GPT Image 2 (as indicated by the author).
2. Copy the template below and replace [CITY] with your city (Barcelona, London, Dubai...).
3. For a thread like TechieSA, repeat the prompt four times with different cities.
4. Keep "1080x1080" if you want a square format for networks.

Prompt to copy

Vertical composition, 1080x1080. A miniature [CITY] made entirely in knitted yarn-art style. Iconic landmarks and architecture of the city crafted in soft yarn with glowing warm interiors. Knitted buildings, yarn-textured streets, yarn-people walking around. Cozy wool textures, handcrafted diorama look, shallow depth of field, highly detailed knitted miniature world of the city.

Template taken from TechieSA response. In English according to the original post.

Four cities of the thread

TechieSA published the same structure for Paris, NYC, Venice and Tokyo. Each carousel image shows recognizable landmarks reinterpreted in yarn: buildings with visible warm interiors, yarn-textured streets, and strolling yarn-people figurines. The finish points to a handmade diorama with reduced depth of field, not a cold 3D render.

What makes this prompt special

The formula is short but precise: miniature + yarn-art + iconic landmarks + bright interiors + knitted people. "Glowing warm interiors" provides a cozy contrast to the matte wool exterior; "handcrafted diorama look" and "shallow depth of field" reinforce the feeling of a model photographed on a table, very shareable on Instagram and Pinterest.

Being a template with a single variable ([CITY]), it scales well: one city per image or a 2×2 collage asking for four cities in a single prompt — although TechieSA chose to generate four separate files at 1080×1080.

City ideas

• Barcelona → Sagrada Familia + woven ramblas
• London → Big Ben + double decker bus in yarn
• Dubai → Burj Khalifa + golden thread facades
• Rome → Colosseum + woolen cobbled streets

Before publishing

The generated landmarks can approximate protected brands or buildings; do not use results with recognizable logos in commercial campaigns without review. Respect OpenAI and SocialSight policies for image generation.