Loading
Vous avez un workflow ComfyUI ? Obtenez une documentation lisible directement dans votre navigateur.Rendre un workflow lisible.Essayer maintenant ↗
V67 Perfect Prompt LoRA for Flux
Source du modèle
Description de la source
V67 Perfect Prompt LoRA for Flux
Premise
If you decide to use my prompt engineering method, I will give you 500 Yellow Buzz if your image (generated using these techniques) is chosen for the cover of CIVITAI , even if combined with other models.
R+ images will be hidden from the gallery. This methodology is model-agnostic but optimized for Flux.
Putting a Like to the V67 models is worth more than giving me Buzz (yellow/green).
Sources
1 sourceArtefacts du modèle
1 artefactV67_Perfect_prompt.safetensors
safetensors · 18,4 MB · SHA-256 e4ea91671835…fc90 · Civitai
TéléchargerExtraits de sources
1 extraitSource context: 190 downloads · Type LORA · Base model Flux.1 D
The Art of Words, The Science of Vision
LoRA component name: V67 Perfect Prompt
Try my tools: https://civitai.com/user/angelomaiota/models
Generate perfect prompts here: https://perchance.org/promptyx
Read my article on Joy Caption Beta One (LM Studio edition): https://civitai.red/articles/29197/writing-a-perfect-prompt 👉 The article explains how to install and use LM Studio (PC/Mac) to run Joy Caption Beta One locally – no cloud, full privacy, perfect for batch captioning and prompt reverse-engineering.
Promptyx features
🧩 PromptyX – Complete Features (Single-Column View)
About PromptyX Advanced AI image generation tool that turns simple ideas into stunning, technically optimized prompts and images. It combines AI enhancement, a built-in editor, and an image-to-prompt converter.
Enhance Prompt Rewrites your basic idea into a detailed, professional-grade prompt based on your chosen style, phrasing, and length, then automatically generates images.
Style Selector Choose from styles like Realistic, Anime, Cinematic, Fantasy, Cyberpunk, and more. Each style applies its own expert prompt-engineering rules and visual tags.
Phrasing Selector Controls how the prompt is written – Tags (comma-separated keywords), Anatomical (technical/structural description), Creative (flowing narrative prose), or Technical (camera/render-spec style) – independently from the visual style.
Length (S/M/L) Sets how short or richly detailed the enhanced prompt should be.
Negative Prompt Toggles a field where you can list things to exclude from generated images (e.g., blurry, deformed).
Example Button Loads a random creative example prompt to get you started quickly.
Reset Button Clears the input field, selected style/phrasing, and all results in one go.
Image to Prompt (Camera Button) Opens the Image to Prompt tool – upload or drag & drop any PNG, JPG, or WEBP image.
Analyze & Generate The AI detects the image's own visual style and details, reverse-engineers a matching prompt, and immediately generates new images from it.
Copy Prompt Copies the enhanced prompt text to your clipboard with one click.
Edit Prompt Lets you manually edit the enhanced text; saving it regenerates images with your changes.
AI Smart Edit Describe a change in plain language (e.g., "make it night") and the AI rewrites the prompt for you.
Regenerate Prompt Re-runs the enhancer on your original input to produce a fresh variation.
Version Navigation Step back and forth between previous edits or AI-generated edits of the same prompt.
Generate More Adds additional image variations to the same result card.
View Toggle Switches a result card between horizontal scroll and grid layout.
Delete Removes a result card from the page.
Save to Gallery Saves an individual image to your device's on-device Private Gallery.
Settings Configure how many images are generated per prompt and their aspect ratio (Square, Portrait, Landscape, or Mixed).
Generation History Every enhanced prompt is saved automatically – search, reuse, copy, or delete past entries anytime.
Saved Prompts Every generated prompt is bookmarked here for quick copying and reuse later.
Private Gallery Browse, manage, and revisit every image you've saved on this device.
AI Assistant A quick chat for image ideas or short prompt-improvement suggestions.
What's New Shows the latest updates and features added to PromptyX.
Menu Button Opens the sidebar with all tools, links, and settings.
Trigger Words for V67 Perfect Prompt
(Higher weight than "Perfect Prompt")
Use this line at the beginning of every prompt:
(V67 Perfect Prompt:1.2), Masterpiece, ultra detailed, cinematic, structured, high fidelity, technical precision, narrative depth, complex scene, maximum specificity, prompt engineering
Note: The LoRA component V67 Perfect Prompt is activated by this exact trigger string. The weight 1.2 gives it priority over the generic quality tags that follow.
Overview
Step into the mind of a master prompt engineer. This is not a LoRA – it is a philosophy, a structured approach, and a curated vocabulary designed to generate perfect prompts. Learn how to blend the right words, reference the greatest photographers and painters, specify professional cinematography equipment like the ARRI Alexa, and achieve maximum specificity and narrative depth.
“A perfect prompt is not a wish – it is a blueprint.”
Why Perfect Prompt Engineering Matters
AspectWithout EngineeringWith Perfect Prompt Engineering ClarityVague, ambiguousPrecise, unambiguousStyleGeneric, randomDeliberate (e.g., cinematographer X, painter Y)Technical ControlNoneFull control (lens, aperture, lighting, camera)ReproducibilityLowHighNarrativeAbsentEmbedded, emotional, contextual
The Three Pillars of a Perfect Prompt
Who? What? Doing what?
Example: “A lone samurai kneeling in a rain-soaked bamboo forest, drawing his katana slowly.”
Borrow the eyes of the greats. Combine photographers, painters, and cinematic references.
CategoryNamesUse CasePhotographers Ansel Adams (landscape), Helmut Newton (b&w, dramatic), Steve McCurry (color, portrait), Sebastião Salgado (documentary), Gregory Crewdson (cinematic staged)Realism, lighting, composition Painters Caravaggio (chiaroscuro), Vermeer (window light), Turner (atmosphere), Hopper (isolation, light), Mucha (art nouveau), Rothko (color fields)Mood, color, texture Cinematographers Roger Deakins (naturalistic light), Emmanuel Lubezki (long takes, natural light), Hoyte van Hoytema (texture, grain), Vittorio Storaro (color symbolism)Motion, depth, cinematic feel Directors Kubrick (symmetry), Wes Anderson (color palette), Tarkovsky (dreamlike), Nolan (practical effects)Composition, tone
Professional camera, lens, and lighting data grounds the image in reality.
Cinema cameras (best for Flux):
ARRI Alexa 65 (medium format sensor, shallow depth, cinematic roll-off)
ARRI Alexa Mini LF (versatile, documentary feel)
RED Monstro 8K (clinical sharpness)
Sony Venice (rich colors)
Lenses:
Cooke S4/i (warm, organic)
Zeiss Master Prime (ultra-sharp, neutral)
Panavision Anamorphic (oval bokeh, lens flares)
Helios 44-2 (swirly bokeh, vintage)
Lighting terms:
Chiaroscuro , Rembrandt lighting , butterfly lighting , golden hour , practical lights , diffused softbox , single source key light
Film stock / digital emulation:
Kodak Portra 400 , Fuji Pro 400H , Ilford HP5 (b&w), 35mm grain , medium format , IMAX
The Perfect Prompt Formula
text [Subject description] + [action/emotion] + [environment/background] + [lighting] + [camera & lens] + [film style or master reference] + [quality triggers]
Example structured prompt: An elderly fisherman mending a net, weathered hands, concentrated gaze, sitting on a wooden dock at dawn, golden side light, shot on ARRI Alexa 65 with Cooke S4/i 50mm, shallow depth of field, Kodak Portra 400 emulation, style of Sebastião Salgado, masterpiece, ultra detailed, cinematic.
Trigger Words for Perfect Prompts (Higher Weight than "Perfect Prompt")
Use this line at the beginning of every prompt:
(Perfect Prompt:1.2), Masterpiece, ultra detailed, cinematic, structured, high fidelity, technical precision, narrative depth, complex scene, maximum specificity, prompt engineering
Recommended Settings for Flux (when using these prompts)
ParameterFlux SchnellFlux Dev / Pro LoRA Weight (if any)0.5 - 0.70.6 - 0.8CFG Scale2.0 - 3.53.5 - 5.0Steps4 - 625 - 40SamplerEulerEuler, DPM++ 2M KarrasGuidance2.5 - 3.54.0 - 5.5
Linking Tools for Perfect Prompts
Prompt generator: https://perchance.org/promptyx – use it to experiment with word combinations before writing your final prompt.
Joy Caption Beta One + LM Studio (PC/Mac) – discussed in my Civitai article ( https://civitai.red/articles/29197/writing-a-perfect-prompt ). This is an advanced captioning program that runs locally via LM Studio . Download LM Studio for Windows or macOS, load a compatible LLM, and use Joy Caption Beta One to reverse-engineer images into perfect prompts. No cloud, no API costs, full privacy.
Bookmark both. They are your prompt engineer’s workshop.
10 Example Prompts (Following the Perfect Prompt Methodology)
Each prompt below uses the trigger prefix and integrates masters, photographers, painters, ARRI Alexa, and technical details.
Prompt 1 – Portrait (Caravaggio + Alexa)
(Perfect Prompt:1.2), Masterpiece, ultra detailed, cinematic, structured, high fidelity, technical precision, narrative depth, complex scene, maximum specificity, prompt engineering – A young woman with tear-streaked cheeks, holding a single extinguished candle, dark background, dramatic chiaroscuro lighting from upper left, style of Caravaggio, shot on ARRI Alexa Mini LF with Zeiss Master Prime 85mm at f/1.4, shallow depth of field, soft film grain, melancholy expression, oil painting texture mixed with digital realism.
Prompt 2 – Landscape (Ansel Adams + Turner)
(Perfect Prompt:1.2), Masterpiece, ultra detailed, cinematic, structured, high fidelity, technical precision, narrative depth, complex scene, maximum specificity, prompt engineering – A vast mountain valley at sunrise, mist rising from a river, dramatic clouds catching golden light, foreground jagged rocks, atmosphere of J.M.W. Turner, composition of Ansel Adams, captured on ARRI Alexa 65 with Panavision anamorphic 40mm, f/5.6, warm golden hour light, large format presence, ethereal yet sharp.
Prompt 3 – Street Photography (Steve McCurry + Alexa)
(Perfect Prompt:1.2), Masterpiece, ultra detailed, cinematic, structured, high fidelity, technical precision, narrative depth, complex scene, maximum specificity, prompt engineering – A bustling spice market in Marrakech, elderly vendor with a weathered face and bright turban, laughing, surrounded by colorful pyramids of spices, natural window light from above, style of Steve McCurry, shot on ARRI Alexa LF with Cooke S4/i 35mm, f/2.8, vibrant saturated colors, shallow depth on vendor's face, background slightly blurred, documentary realism.
Prompt 4 – Cinematic Narrative (Roger Deakins + ARRI)
(Perfect Prompt:1.2), Masterpiece, ultra detailed, cinematic, structured, high fidelity, technical precision, narrative depth, complex scene, maximum specificity, prompt engineering – A lone detective in a trench coat standing under a streetlamp in heavy rain, wet asphalt reflecting neon signs, night scene, lighting by Roger Deakins (Blade Runner 2049 style), shot on ARRI Alexa 65 with Panavision anamorphic 50mm, t/2.8, deep focus, cool teal and orange contrast, fine grain, cinematic anamorphic flares.
Prompt 5 – Fine Art Still Life (Vermeer + Mucha)
(Perfect Prompt:1.2), Masterpiece, ultra detailed, cinematic, structured, high fidelity, technical precision, narrative depth, complex scene, maximum specificity, prompt engineering – A vase with sunflowers and wilting peonies on a wooden table, soft northern window light from the left, a blue velvet drape behind, style of Vermeer mixed with Alphonse Mucha’s decorative elegance, shot on ARRI Alexa Mini with Helios 44-2 58mm, f/2, natural light, warm ivory and deep blue tones, painterly brushstroke texture.
Prompt 6 – Monochromatic (Helmut Newton + Ilford HP5)
(Perfect Prompt:1.2), Masterpiece, ultra detailed, cinematic, structured, high fidelity, technical precision, narrative depth, complex scene, maximum specificity, prompt engineering – A fashion model in a sharp tailored suit,...
