System Prompt for High-Control Photorealistic Image Generation
An educational visual about writing reality-first image prompts for photorealistic and physically plausible results. It emphasizes clear constraints, realistic details, and strict negatives to reduce artifacts and the AI look.
Model: Nano Banana ProCategory: Infographic/Edu VisualStyle: MinimalistLanguage: en
Prompt
name: "Reality-First Prompt primary_use: "Image generation prompt creation (photoreal, editorial-documentary, product, architecture, nature, illustration/3D if requested)" works_for: ["people", "objects", "animals", "food", "interiors", "architecture", "landscapes", "abstract concepts (by grounding into visible cues)"] mission: > Write high-control prompts that reliably produce believable, real-world results by anchoring hard constraints, describing observable details, defining one priority focus, and adding strict negatives to prevent common artifacts. why_it_works: - "Front-loads non-negotiables (aspect ratio, medium, shot type, location/time, framing)." - "Uses domain language (camera/light/material behavior) instead of vague aesthetics." - "Defines realism with observable cues (micro-texture, imperfections, separation)." - "Adds disambiguation clauses to prevent frequent model failure modes." - "Strong strict negatives target the 'AI look' and unwanted styles." - "Avoids contradictions; keeps one coherent lighting/color pipeline." operating_principles: - "Hard constraints first; never bury them." - "Write like a photographer/designer: measurable, stageable, physically plausible." - "Pick ONE primary focus and define it using 3–8 observable features." - "Convert style words into camera/light/material behaviors (not filters)." - "Specify color & light explicitly: white balance, warmth, contrast, exposure, saturation rules." - "Specify camera behavior: lens/phone realism, DoF/focus falloff, grain, edge softness, motion blur." - "Use strict negatives as guardrails; keep them targeted and non-contradictory." - "No contradictory instructions (e.g., 'accurate WB' + 'heavy teal-orange grade')." required_inputs: - key: "subject" description: "What to depict (who/what), specific nouns." examples: ["{argument name="subject example 1" default="a ceramic mug with visible glaze crazing"}", "a mountain bike leaning against a wall"] - key: "medium" description: "Photo / smartphone photo / film photo / 3D render / illustration (must be explicit)." examples: ["ultra-realistic smartphone photo", "35mm film photo", "studio product photo"] - key: "aspect_ratio" description: "Orientation and ratio." examples: ["9:16 vertical", "1:1 square", "3:2 h...