
2x2 grid, 16:9, 4 days that changed history: RENDER_CHAIN = "[$SUBJECT :: flow_strips :: nested_quilling_sculpture :: raking_macro]" FUSION_RULES = { "flow_analysis": "compute edge-direction field θ and curvature field κ across $ SUBJECT; derive flow vector field F = ∇θ; curl(F) drives coil tightness, divergence(F) drives coil spacing", "strip_conversion": "convert every streamline of F into a tightly coiled quilling strip; coil radius = 1/|κ| at that point; strip follows the flow without breaking it", "recursion": "each large coil recursively contains 2–4 smaller nested coils aligned to the same local flow direction; recursion depth capped so the silhouette stays readable", "material": "dyed paper quilling strips, satin finish, low specular, visible fibrous edges; gilded tips (gold leaf) at terminal ends of strips only; no other material", "support": "shallow white plinth, no visible armature, no wire, no plastic", "lighting": "global illumination + soft top fill; hard raking key from ~20° above horizon, aligned to F; reveals micro-relief on every coil edge and paper fiber", "camera": "shallow DOF f/1.8; focal plane on the nearest coil edge cluster; paper fibers resolved at the focal plane; background coils fall into soft bokeh", "result": "a quilled-paper sculpture of $ SUBJECT in which every curve becomes nested spirals of razor-clean paper geometry", "negative": "wire armature, plastic, glossy resin, painted illustration, flat 2D, 3D render sheen, blurred paper fibers, random coils off-flow, pasted-on decoration" } OUTPUT = (RENDER_CHAIN :: FUSION_RULES) - NEGATIVE_SET RENDER: ONE macro photograph of the sculpture.
This image was generated with Nano Banana 2.1 by Google. The prompt on this page is the exact text used to create it.
The example was generated with Nano Banana 2.1 using: 120:67 · Surreal · Abstract · Macro. Reusing these settings gives the closest match.
Copy the prompt as-is first, then replace only the subject while keeping the style, lighting, and composition keywords. Generate a few variations — outputs differ between runs.
Yes. Prompts work best as starting points — adjust descriptors, aspect ratio, or style tags to make the output your own.