Prompting Cheat Sheets

• Quality tags first — "RAW photo, photorealistic, 8K"
• Subject — who/what with full detail: age, clothing, expression
• Lighting — golden hour, Rembrandt, softbox, neon practical
• Camera — "Canon EOS R5, 85mm f/1.2" = photorealism signal
• Mood/Style — "editorial, serene, Vogue quality"
• Negative prompt — always include, never skip

• For portraits add: airbrushed, asymmetrical face
• For products add: hands, background clutter

• Golden hour — warm 5600K backlight. Portraits and outdoor editorial
• Rembrandt — dramatic triangle shadow. Character studies
• Studio softbox — clean, even. Product shots and beauty
• Neon / practical — colorful atmospheric. Urban, night scenes
• Overcast natural — diffused, shadow-free. Documentary
• Blue hour — cool ambient sky. Architectural
• Window light — soft directional. Indoor editorial

• Canon EOS R5, 85mm f/1.2 — portrait, creamy bokeh
• Sony A7R IV, 35mm f/2 — street and candid
• Canon EOS R5, 24mm f/8 — wide landscape, everything sharp
• Hasselblad H6D, 120mm f/16 — product/macro, zero distortion
• Leica M11, 50mm f/1.4 — film aesthetic, timeless
• ARRI ALEXA, anamorphic — cinema grade, oval bokeh

• ChatGPT Image 2.0 — iterative editing, text in images, conversational workflow
• Seedream 5.0 — cinematic scenes, complex compositions, high realism
• Midjourney v7 — editorial aesthetics, artistic composition
• Nano Banana Pro 2 — rapid testing
• Ideogram v2 — typography and graphic design

• CFG Scale — how strictly AI follows prompt. 4–6 = sweet spot
• Steps — refinement passes. 20–28 is quality ceiling
• Seed — same seed + same prompt = same image. Lock before iterating
• Denoising strength — img2img: 0.35–0.50 = enhance, preserve
• Prompt weighting — (word:1.4) to boost, (word:0.5) to reduce
• LoRA — add-on file that specializes a model for a style or face

• Start with motion — describe what is moving first
• Verb-forward language — "walks slowly" beats "is walking"
• Lock your subject — "same face throughout, no morphing"
• Specify camera — tracking shot, dolly, aerial, handheld
• Name a director — Roger Deakins, Denis Villeneuve
• Set duration — "8 seconds, 24fps, no scene cuts"

• Seedance 2.0 — human motion, dance, sports
• Higgsfield Studio — upload face, generate that person in video
• Kling 3.0 — physics-accurate action, water, fire, fabric
• Runway Gen-4 — VFX, transformations, commercial quality — View Runway Gen-4 guide →
• Pika — stylized, animated, artistic content
• Agent Opus — auto-edits video into TikTok/Reels clips

• Tracking dolly — moves alongside subject
• Dolly push-in — slow move toward subject, dramatic
• Handheld — natural shake, documentary feel
• Aerial drone descending — establishing wide shot
• Locked off tripod — static, subject moves within frame
• Pan — rotates on fixed axis

• "same consistent face and hair throughout all frames"
• "no morphing or flickering between frames"
• "no scene cuts" — forces single continuous clip
• Keep under 120 words — models ignore the rest
• Upload face photo to Higgsfield for identity fidelity

• Roger Deakins — natural practical lighting, restrained color
• Denis Villeneuve — vast scale, epic atmosphere
• Wong Kar-wai — warm neon, slow motion, melancholic
• Christopher Doyle — vibrant colors, fluid handheld
• Sofia Coppola — soft natural light, pastel palette
• Ridley Scott — atmospheric environments, epic scale
• David Fincher — desaturated, precise, clinical tension

• Role — "You are a senior brand copywriter for Chanel..."
• Context — "...writing for affluent women 30-50..."
• Task — "...write a 120-word product description."

• Forces model to reason before responding
• Best for: analysis, strategy, multi-step decisions
• Works in ChatGPT, Claude, Gemini
• Reduces errors on complex problems significantly

• Specify format: "respond in JSON", "max 150 words"
• Exclude explicitly: "no corporate jargon", "no bullet points"
• Few-shot: show 2-3 example outputs you want matched
• XML tags: "wrap answer in <answer> tags"
• Word limits: forces concision, removes padding

• System prompt — instructions set before conversation begins
• Zero-shot — no examples provided
• Few-shot — 2-5 examples before your request
• Temperature — 0 = predictable, 1 = creative/varied
• Context window — how much the AI can "remember"
• Hallucination — AI confidently states incorrect info

AI influencers are fully computer-generated personas — they post content, promote products, and build audiences without ever physically existing. Lil Miquela has 3M+ followers and deals with Prada and Calvin Klein.

• 24/7 availability — never tired or unavailable
• Zero talent fees — no agents, no contracts
• Full brand control — persona never goes off-message
• Global scalability — same face, every market

• Script — GPT-4o, Claude for screenplay
• Storyboard — ChatGPT Image 2.0 for visual boards
• Characters — APOB AI, Higgsfield Studio
• Scenes — Seedance 2.0, Runway, Kling 3.0
• Voice — ElevenLabs for realistic AI voices
• Music — Suno AI, Udio for original scores
• Edit — Agent Opus for automated assembly

• Step 1 — Create persona: APOB AI or Glam AI
• Step 2 — Photos: ChatGPT Image 2.0 with reference
• Step 3 — Video: Higgsfield Studio
• Step 4 — Captions: GPT-4o brand voice prompt
• Step 5 — Short clips: Agent Opus
• Step 6 — Automate with n8n pipeline

• Define your niche — fashion, fitness, travel, tech, luxury
• 3 posts/week — consistency over volume
• Platform strategy — Instagram for photos, TikTok for video
• Brand voice — use GPT-4o brand voice system prompt for all copy
• Brand deals — pitch brands at 10K followers
• Affiliate income — promote AI tools for passive revenue

AI models weight earlier tokens more heavily. Put your subject, role, or most critical instruction first — not at the end. "1woman, golden hour, Canon 85mm" outperforms "Canon 85mm golden hour 1woman" because the subject is anchored first.

Write "1woman" not "a woman." Writing "1" prevents the model from generating groups, doubles, or ambiguous compositions. Works for any subject: 1man, 1cat, 1product. Explicit numbers = explicit intent.

"Shot on Canon EOS R5, 85mm f/1.2" is a quality signal. The model has learned that images tagged with professional cameras are high quality — naming the camera pulls the output toward that quality level automatically. Use it every time.

"Vogue Japan editorial" or "Kinfolk magazine aesthetic" communicates an entire visual world in two words. Publication references invoke lighting style, color palette, composition rules, and mood all at once. More precise than describing each element individually.

Negative prompts are not optional. Without them the model defaults to its highest-probability output — which is usually plastic skin, flat lighting, and stock photo energy. Always add: CGI, plastic skin, airbrushed, watermark, blurry, overexposed. 30 seconds. Significant quality improvement.

Not locking subject identity. Without explicit instructions, the model drifts — face, hair, and clothing change between frames. Always add: "same consistent face and hair throughout every frame, no morphing, no identity change." For Higgsfield: use a reference photo as the identity anchor.

As long as it needs to be — no longer. A 20-word prompt with specific, well-chosen elements outperforms a 150-word prompt full of filler. For image prompts: 40–80 words is the sweet spot. For video: 60–120 words. For LLMs: as long as the context requires, but structured clearly.

AI models use a random seed by default. The seed controls the starting noise — same prompt, different seed, different output. To reproduce a result you like, save the seed number from the output metadata and lock it for subsequent runs.

Yes, significantly. Models weight earlier tokens more heavily in most architectures. Subject first, then lighting, then camera, then style, then quality tags. For LLMs: role definition first, then context, then task, then constraints. Front-load what matters most.

For photorealism: CFG 5–7. For artistic/stylized: CFG 7–9. Above 9 risks over-saturation and artifacts. FLUX models prefer lower CFG (3–5) than SDXL-based models (6–8). ChatGPT Image 2.0 handles CFG internally — no manual setting needed.

20 steps for fast drafts. 25–30 for final production output. Above 30 produces diminishing returns and significantly increases generation time. The sampler matters more than step count above 25 — DPM++ 2M Karras at 25 steps consistently outperforms Euler at 50 steps.

Token limit and weighting. Very long prompts cause earlier elements to lose influence. Conflicting instructions cancel each other out. The model has trained on millions of images — some concepts are so rare in training data they're effectively invisible to the model. Simplify and front-load the critical elements.

In SDXL and FLUX pipelines: use (word:1.3) to increase emphasis on a term, (word:0.8) to reduce it. Example: "(golden hour:1.4), (Canon 85mm:1.2), (natural skin:1.3)" tells the model which elements matter most. Don't overweight — above 1.5 causes artifacts.

Generate a base image → img2img at 0.40 denoising to refine → img2img at 0.35 for final polish. Each pass preserves the composition while improving detail. The Universal 8K Upscale prompt is specifically built for the final pass. Three-pass chain = commercial-grade output from any starting image.

A reference image communicates specificity that 200 words can't. For face consistency: use IP-Adapter Face ID or Higgsfield's reference upload. For style: use IP-Adapter at 0.6–0.7 weight. For video: a reference photo in Higgsfield Studio locks identity across every frame more reliably than any text instruction.

Run your rough idea through the Meta-Prompt Upgrader (Prompt 22 in the vault) before sending it to an image or video model. The LLM adds camera specs, lighting, mood direction, and negative prompts automatically. Takes 30 seconds. Measurably improves output quality on every run.

Find a seed that produces a subject with the exact features you want. Lock the seed. Vary only the environment, lighting, and outfit across generations. This technique produces a consistent "character" across an entire image series without ControlNet or reference images — using only the seed as the identity anchor.

For complex LLM tasks, add "Think step by step before answering." This forces the model to surface its reasoning before committing to an output — catching errors it would otherwise skip past. Dramatically improves accuracy on multi-step reasoning, planning, and analysis tasks across all major LLMs.