You are a generative art prompt engineer specializing in NFT collection production and AI art consistency. Create production prompt system for: My Collection Specs: - Collection size: [number of NFTs] - AI tool: [Midjourney v6, DALL-E 3, Stable Diffusion, etc] - Subject: [characters, objects, abstract, etc] - Trait categories: [how many trait types] - Rarity distribution: [common, uncommon, rare, legendary %] - Timeline: [generation schedule] Develop complete prompt engineering system: 1. Modular Prompt Architecture Base Prompt Components: Core Structure: [Subject] + [Style] + [Attributes] + [Environment] + [Technical] + [Constraints] Subject Module (consistent): "A [character type] [species/form]" Stays mostly the same across collection Style Module (consistent): "in [art style], [color palette], [lighting style], [texture]" Collection's signature look Attributes Module (variable): "with [trait 1], [trait 2], [trait 3], wearing [clothing], holding [item]" Where variation happens Environment Module (variable): "[background type], [setting], [atmosphere]" Context and rarity differentiation Technical Module (consistent): "digital art, highly detailed, 8k, professional quality" Quality standards Constraints Module (consistent): "--no text, watermarks, logos, multiple subjects, distortion" What to avoid Generate 10 complete modular prompts showing system in action. 2. Trait Variation System Trait Category Framework: Category 1: [Name] (Example: Headwear) Common variants (70%): - Trait 1: "wearing [item 1]" - Trait 2: "wearing [item 2]" - Trait 3: "wearing [item 3]" Uncommon variants (20%): - Trait 4: "wearing [special item 1]" - Trait 5: "wearing [special item 2]" Rare variants (8%): - Trait 6: "wearing [rare item]" Legendary variants (2%): - Trait 7: "wearing [legendary item]" Repeat for all trait categories. Generate complete trait prompt library for 5-10 categories with 20-30 total traits. Trait Combination Rules: Compatible Combinations: - [Trait A] works with [Trait B, C, D] - [Trait E] conflicts with [Trait F] - [Trait G] requires [Trait H] Rarity Stacking: - Maximum rare traits per NFT: [number] - Legendary traits: [standalone or combinable] - Rarity multipliers: [how traits affect overall rarity] Create combination matrix preventing visual conflicts. 3. Batch Generation Workflow Production Process: Step 1: Trait Assignment (spreadsheet): NFT #1: [Trait combo generated randomly by rarity %] NFT #2: [Different combo] [Continue for full collection] Step 2: Prompt Assembly: Base + Trait 1 + Trait 2 + Trait 3... = Complete Prompt Step 3: Generation: - Feed prompts to AI tool - Use consistent settings/seeds when possible - Generate multiple variations - Select best of 4-6 outputs Step 4: Quality Control: - Check style consistency - Verify traits match assignment - Flag issues for regeneration - Approve for collection Create detailed workflow document with timing estimates. Daily Generation Schedule: Day 1: Generate 50 NFTs (Batch A) - Morning: Prompts 1-25 - Afternoon: Prompts 26-50 - Evening: QC and selection Day 2: Generate 50 NFTs (Batch B) [Continue until collection complete] For 1000 NFT collection at 50/day = 20 days generation time Build production calendar with milestones. 4. Platform-Specific Optimization Midjourney Prompt Structure: Format: /imagine prompt: [full prompt] --ar 1:1 --style raw --v 6 --s 250 Midjourney-Specific Tips: - Use "--s" (stylize) for consistency - "--chaos" for variation control - "--weird" for creative elements - Multi-prompts for complex scenes: [element 1]:: [element 2]:: - Image weights: [reference image] --iw 0.5 Generate 20 Midjourney-optimized prompts with parameters. DALL-E 3 Prompt Structure: Format: [Detailed description], [style], [technical quality], [specific instructions] DALL-E-Specific Tips: - Very detailed descriptions work better - Natural language over keywords - Specific about what you DON'T want - Less technical jargon, more descriptive - Aspect ratio in natural language Generate 20 DALL-E 3 optimized prompts. Stable Diffusion Prompt Structure: Format: [Subject], [style keywords], [quality tags] Negative prompt: [everything to avoid] Settings: - Steps: [20-30 for quality] - CFG Scale: [7-12 for prompt adherence] - Sampler: [Euler a, DPM++, etc] - Model: [checkpoint specific to your style] Generate 20 Stable Diffusion prompts with settings. 5. Consistency Maintenance Techniques Style Anchor Methods: Seed Locking: - Use same seed for trait families - Document successful seeds - Seed ranges by rarity tier Reference Images: - Create 10 "style master" images - Use img2img at 30-50% strength - Maintains visual DNA ControlNet (Stable Diffusion): - Lock pose/composition - Ensure character consistency - Depth maps for environment Style Training: - Fine-tune model on your art (advanced) - DreamBooth or LoRA training - Ultimate consistency Create consistency checklist for each method. 6. Quality Control Framework Acceptance Criteria: Technical Quality: - Resolution: [minimum pixels] - No artifacts or glitches - Clean edges - Proper rendering Style Consistency: - Matches collection aesthetic - Color palette adherence - Recognizable as part of series - No style drift Trait Accuracy: - All assigned traits present - Traits rendered correctly - No conflicting elements - Rarity appropriately reflected Uniqueness: - Doesn't duplicate existing NFT - Distinct composition - Adds value to collection Create 50-point QC checklist for every generated piece. Rejection and Regeneration: Common Rejection Reasons: - Style drift from collection - Missing or incorrect traits - Technical quality issues - Too similar to another NFT - Composition problems Regeneration Strategy: - Adjust prompt specificity - Modify seed or parameters - Try different variations - Use img2img refinement - Last resort: manual editing Document rejection rate targets (aim for <20% rejection). 7. Rarity Visual Hierarchy Visual Differentiation by Tier: Common (70%): Prompt additions: "simple [element], standard [feature]" Visual: Clean, straightforward, good quality Uncommon (20%): Prompt additions: "detailed [element], enhanced [feature], [special effect]" Visual: More elaborate, eye-catching details Rare (8%): Prompt additions: "highly detailed [element], glowing [effect], dynamic [feature], cinematic lighting" Visual: Clearly special, multiple effects Legendary (2%): Prompt additions: "masterpiece quality, epic [element], multiple [effects], dramatic lighting, ethereal glow, particle effects, [unique composition]" Visual: Unmistakably rare, jaw-dropping Generate prompt modifiers for each rarity showing escalation. 8. Troubleshooting Common Issues Problem-Solution Matrix: Issue: Inconsistent character features Solution: Add more specific anatomical details to base prompt Fix: "symmetrical face, consistent proportions, [specific feature details]" Issue: Colors shifting between generations Solution: Lock color palette with hex codes Fix: "color palette of #[hex1], #[hex2], #[hex3], cohesive color scheme" Issue: Traits not rendering correctly Solution: Increase prompt weight or reorder Fix: Move trait earlier in prompt, add emphasis: "(red hat:1.3)" Issue: Style drift over time Solution: Regenerate style anchors, use reference images Fix: Start fresh session with style masters loaded Issue: Background inconsistency Solution: Standardize background prompts Fix: Create 5-10 background templates, use consistently Create 20-problem troubleshooting guide with solutions. 9. Advanced Prompt Techniques Prompt Weighting (Stable Diffusion/Midjourney): (keyword:1.2) = emphasize (keyword:0.8) = de-emphasize Example: "(blue eyes:1.3), (red jacket:1.2), (background:0.8)" Multi-Prompt Composition (Midjourney): [main subject]:: [background]:: [effects]::2 Numbers control relative importance Negative Prompting: Stable Diffusion: Separate negative prompt field Midjourney: --no keyword DALL-E: "without [things]" Generate advanced prompt examples using these techniques. 10. Production Optimization Speed vs Quality Balance: Fast Generation (for testing): - Lower steps/iterations - Batch processing - Accept more variations - 100+ NFTs per day possible Quality Generation (for final): - Higher steps/iterations - Careful selection - Multiple variations per prompt - 20-50 NFTs per day realistic Create strategy for each phase of production. Automation Opportunities: Prompt Assembly: - Spreadsheet formulas combine modules - Python scripts generate prompts - Template systems - Bulk processing Batch Submission: - Queue systems - Overnight generation - API integration (where available) - Monitoring automation Post-Processing: - Automated upscaling - Batch cropping/formatting - Watermark removal - File organization Generate automation workflow for 1000+ NFT collection. Write everything practically: - Platform-specific instructions - Copy-paste ready prompts - Step-by-step processes - Troubleshooting solutions Provide complete materials: - 200 prompt examples by category - Modular prompt builder template - Trait combination spreadsheet - QC checklist (50 points) - Production timeline calculator - Troubleshooting guide Create complete production system that generates thousands of unique, cohesive, collection-quality NFTs efficiently and consistently.