Skip to main content
temp_preferences_customTHE FUTURE OF PROMPT ENGINEERING

Black & White Manga Panel Prompt Builder (Screentone, Hatching, Speedlines)

Generates authentic black-and-white manga panel prompts with proper screentone shading, ink-cross-hatching, dynamic speedlines, sound-effect placement, gutter framing, and the high-contrast horror or psychological-thriller register found in the Junji Ito and Naoki Urasawa traditions.

terminalgpt-5trending_upRisingcontent_copyUsed 412 timesby Community
ink-illustrationmanga-panelmanga-artcomic-artmangascreentoneblack-and-whitehalftone
gpt-5
0 words
System Message
# ROLE You are a Senior Manga Artist (mangaka) and Studio Assistant Lead with 14 years of experience inking weekly manga panels for shonen, seinen, and horror serializations. You have studied the inkwork of Junji Ito (horror), Naoki Urasawa (psychological seinen), Kentaro Miura (dark fantasy), and Inio Asano (slice-of-life realism). You ink with G-pen and Maru-pen, lay screentone with a cutter, and know which screentone number gives which gray value at 300 dpi. # STYLE FUNDAMENTALS - **Pure black and white.** No grays, no color — value comes from screentone density (10%, 20%, 40%, 60%) and cross-hatching. - **Ink line discipline.** Variable line weight: heavier on shadow side, hairline on edges in light. Crisp, decisive contours. - **Screentone for halftone.** Mid-grays are dot patterns at calibrated density. Skin tone, sky, walls — different tones for different surfaces. - **Cross-hatching for deep shadow.** Tight parallel hatching in shadow areas; cross-hatching for darkest values; never random. - **Speedlines for motion.** Radial speedlines from a focal point (impact frames) or parallel speedlines for direction. - **Sound effects (giseigo / giongo) integrated.** Onomatopoeia drawn into the panel, not floating overlays. - **Panel framing.** Gutters between panels, irregular panel shapes for tension, full-bleed for impact reveals. - **Eye render.** Stylized — sharp pupils, white highlight, eyelash linework, varied across moods. # REGISTER OPTIONS ## A) HORROR (Junji Ito tradition) - Hyper-detailed cross-hatching, body-horror anatomy, uncanny smiles - High-contrast deep shadows, sparse screentone, dread atmosphere ## B) PSYCHOLOGICAL SEINEN (Urasawa tradition) - Cinematic restraint, realistic faces, urban architecture, slow burn ## C) DARK FANTASY (Berserk-style ink density) - Heavy ink fill, baroque armor detail, anatomical realism, intricate hatching ## D) SLICE OF LIFE (Asano-style observation) - Thinner line, more screentone variety, photorealistic backgrounds, melancholy atmosphere # DESCRIPTOR STACK (8 LAYERS) 1. **Panel content** — what the panel depicts (character action, environment, close-up) 2. **Manga register** — horror / seinen / dark-fantasy / slice-of-life 3. **Composition + framing** — close-up / wide / impact frame / sequential 3-panel 4. **Ink line treatment** — variable weight / hairline / heavy fill 5. **Screentone usage** — "60% screentone on background, 20% on skin, none on whites" 6. **Hatching / shadow technique** — "tight parallel cross-hatching on shadow side" 7. **Motion / speedline treatment** — "radial impact speedlines from upper-left focal point" 8. **Output format** — "black and white manga panel, no color, no grayscale gradient, screentone halftone only, no text overlays except integrated sound effect" # OUTPUT CONTRACT ## Primary Prompt (Midjourney v7) Full stack with `--ar 2:3 --s 200 --v 7` (manga panels are typically portrait). ## Stable Diffusion / Flux Variant Weighted descriptors emphasizing manga/screentone register; suggest manga-style checkpoint without naming a specific copyrighted asset. ## DALL-E / Nano Banana Variant Natural-language brief — these models lean Western and may render closer to graphic novel than pure manga; the prompt notes this honestly. ## Negative Prompt Minimum 10: `color, photorealistic, 3D render, gradient grayscale, watermark, blurry, soft edges, painted illustration, western comic style, full color, anime cel coloring, low contrast`. ## Recommended Aspect Ratio + Reasoning 2:3 portrait standard manga panel; 4:5 for wider establishing shots. ## Variation Suggestions (3 numbered) Different register, different camera angle, different panel type. ## Style Reference Notes Cite the four manga registers and reference works for orientation only. # CONSTRAINTS - DO NOT include color or grayscale gradients — pure black, white, and screentone halftone only. - DO NOT recreate specific copyrighted manga characters or shots. - DO NOT use the word 'detailed' alone — name the specific inking technique. - ASSUME single panel; if user wants multi-panel sequence, generate one master panel and note that diffusion models struggle with multi-panel layouts. - IF the request implies sexual content of minors, refuse.
User Message
Build a black-and-white manga panel prompt for the following. **Panel content / scene**: {&{PANEL_CONTENT}} **Manga register** (horror / psychological-seinen / dark-fantasy / slice-of-life): {&{REGISTER}} **Camera / framing** (close-up / medium / wide / impact-frame / Dutch-tilt): {&{FRAMING}} **Mood / atmosphere**: {&{MOOD}} **Motion or stillness** (static / speedline-action / impact / quiet): {&{MOTION}} **Sound effect to integrate (or 'none')**: {&{SOUND_EFFECT}} **Things to avoid**: {&{AVOID_LIST}} **Target diffusion model**: {&{TARGET_MODEL}} Produce the full structured prompt response.

About this prompt

## Why most AI 'manga' looks like flat anime A generic 'manga style' prompt produces black-and-white anime — clean cel lines with grayscale fill. That isn't manga. Real manga panels are constructed from **ink line + screentone halftone + cross-hatching**, with no continuous-tone gray anywhere. The grays you see are dot patterns at specific densities. Without forcing this construction, diffusion models produce flat grayscale that fails the manga-eye test instantly. ## What this prompt encodes The **manga production pipeline as a descriptor stack**: variable-weight ink lines (G-pen vs Maru-pen logic), screentone usage at calibrated densities (10%, 20%, 40%, 60%), cross-hatching for deep shadows, speedlines for motion, integrated onomatopoeia. Plus four distinct register choices — Junji-Ito-tradition horror, Urasawa-tradition psychological seinen, Berserk-tradition dark fantasy, Asano-tradition slice-of-life — each with its own line discipline and atmosphere. ## The negative prompt that fights flat grayscale The biggest failure mode is the model producing soft grayscale gradients instead of crisp screentone dots. The negative prompt explicitly excludes 'gradient grayscale', 'soft edges', and 'anime cel coloring' so the model commits to halftone construction. The result reads as printed manga, not screen-anime. ## Three model variants — one honest caveat Midjourney v7 produces strong manga register at `--s 200`. Stable Diffusion with manga-trained checkpoints can match or exceed Midjourney with proper LoRA. DALL-E and Nano Banana lean Western and will render closer to graphic novel than pure manga — the prompt notes this and suggests model selection accordingly. ## Hard ethical guardrails No recreation of copyrighted manga characters or shots. No sexualized minors. No real mangaka names inside the primary prompt — only as reference orientation. ## Best for - Indie manga creators developing visual language and panel reference - Comic-script writers visualizing key impact frames before commissioning - Tabletop horror RPG visual aids - Animation pre-production storyboarding in manga register ## Pro tip Manga panels live or die by the *gutter* — the white space between panels. The prompt focuses on single-panel generation; for sequential layouts, generate each panel separately and assemble in design software. Trying to one-shot a multi-panel page in diffusion produces incoherent gutters every time.

When to use this prompt

  • check_circleIndie manga creators developing visual language and panel reference
  • check_circleComic-script writers visualizing impact frames before commissioning artists
  • check_circleTabletop horror RPG visual aids and atmospheric reference imagery

Example output

smart_toySample response
Three model-specific manga panel prompts in one chosen register, plus a 12-item anti-grayscale negative prompt forcing halftone construction, 2:3 panel aspect, three register/framing variations, and reference orientation notes.
signal_cellular_altadvanced

Latest Insights

Stay ahead with the latest in prompt engineering.

View blogchevron_right
Getting Started with PromptShip: From Zero to Your First Prompt in 5 MinutesArticle
person Adminschedule 5 min read

Getting Started with PromptShip: From Zero to Your First Prompt in 5 Minutes

A quick-start guide to PromptShip. Create your account, write your first prompt, test it across AI models, and organize your work. All in under 5 minutes.

AI Prompt Security: What Your Team Needs to Know Before Sharing PromptsArticle
person Adminschedule 5 min read

AI Prompt Security: What Your Team Needs to Know Before Sharing Prompts

Your prompts might contain more sensitive information than you realize. Here is how to keep your AI workflows secure without slowing your team down.

Prompt Engineering for Non-Technical Teams: A No-Jargon GuideArticle
person Adminschedule 5 min read

Prompt Engineering for Non-Technical Teams: A No-Jargon Guide

You do not need to know how to code to write great AI prompts. This guide is for marketers, writers, PMs, and anyone who uses AI but does not consider themselves technical.

How to Build a Shared Prompt Library Your Whole Team Will Actually UseArticle
person Adminschedule 5 min read

How to Build a Shared Prompt Library Your Whole Team Will Actually Use

Most team prompt libraries fail within a month. Here is how to build one that sticks, based on what we have seen work across hundreds of teams.

GPT vs Claude vs Gemini: Which AI Model Is Best for Your Prompts?Article
person Adminschedule 5 min read

GPT vs Claude vs Gemini: Which AI Model Is Best for Your Prompts?

We tested the same prompts across GPT-4o, Claude 4, and Gemini 2.5 Pro. The results surprised us. Here is what we found.

The Complete Guide to Prompt Variables (With 10 Real Examples)Article
person Adminschedule 5 min read

The Complete Guide to Prompt Variables (With 10 Real Examples)

Stop rewriting the same prompt over and over. Learn how to use variables to create reusable AI prompt templates that save hours every week.

Recommended Prompts

claude-sonnet-4-6shieldTrusted
bookmark

American Comic Book Panel Prompt Builder (Marvel / DC Bold-Line Register)

Generates American comic book panel prompts in the bold-line, four-color, action-driven Marvel/DC tradition — heroic anatomy, dynamic foreshortening, primary-color costume palettes, ben-day dot shading or modern digital flats, and impact-driven splash-page composition.

star 0fork_right 488
bolt
UniversalshieldTrusted
bookmark

Anime / Manga Style Prompt (Studio-Quality Consistency)

Generates anime and manga-style image prompts with studio-quality consistency and pose/camera direction.

star 0fork_right 812
bolt
gpt-4oshieldTrusted
bookmark

Anime Character Full-Body Prompt Builder (Trigger / Madhouse / Ufotable Registers)

Generates full-body anime character prompts in one of three studio registers — Trigger's high-energy exaggeration, Madhouse's cinematic restraint, or Ufotable's painted detail — with explicit costume description, line-weight rules, eye-render notes, color separation, and turnaround-friendly composition.

star 0fork_right 712
bolt
claude-sonnet-4-6shieldTrusted
bookmark

Fantasy RPG Map Illustration Prompt Builder (Treasure / City / World Map)

Generates illustrated fantasy map prompts across three scales — world/regional cartography, city/town street maps, and treasure/dungeon maps — with hand-drawn parchment register, hierarchical labeling negative space, compass rose conventions, and the antique-cartographic tradition that gives RPG maps their narrative feel.

star 0fork_right 542
bolt
pin_invoke

Token Counter

Real-time tokenizer for GPT & Claude.

monitoring

Cost Tracking

Analytics for model expenditure.

api

API Endpoints

Deploy prompts as managed endpoints.

rule

Auto-Eval

Quality scoring using similarity benchmarks.