chinese gongbi · court painting
နန်းတွင်း ဇင် ပုံပြင်များ
တရုတ် ဂုံပီ ပန်းချီ ပုံစံ · မြန်မာ အသံဖတ် — One-Click Flow
တရုတ် ဇင် (Chan)/ဝူရှား ဒဏ္ဍာရီများကို Qing ခေတ် နန်းတွင်း ဂုံပီ ပန်းချီ ပုံစံဖြင့် ပုံ ၁၀ ပုံ၊ မြန်မာ အသံဖတ် နှင့် ဗီဒီယို prompt များအထိ အလိုအလျောက် ထုတ်ပေးမယ့် Master Prompt တစ်ခုလုံးကို အောက်က တံဆိပ်ခလုတ်ကို နှိပ်ရုံဖြင့် ကူးယူနိုင်ပါသည်။
ယခု Content pack ကို ကျနော် သူရိန်နွေစိုး မှ ကျနော့်သင်တန်းက သူငယ်ချင်းတွေအတွက် ဖန်တီးပေးထားပါသည် ....။
master prompt
နန်းတွင်း ဇင် ပုံပြင် — Content Pack Prompt
အောက်ကစာသားတစ်ခုလုံးကို ကူးယူပြီး Manus ထဲ paste လုပ်ပါ။ ပြီးလျှင် [TITLE] နေရာမှာ ပုံပြင်ခေါင်းစဉ် ထည့်ပါ။
MASTER PROMPT — Chinese Zen Folktale Auto-Production (One-Click, Manus-Ready) TASK: Given only a [TITLE], autonomously produce a complete story production package and deliver it as a single downloadable ZIP file. Do NOT ask for confirmation between steps — execute the full pipeline end-to-end automatically. INPUT: Title: [......] If no title is provided, stop and ask only for the title. Otherwise, proceed automatically through all steps below without pausing. ================================================================ STEP 1 — WRITE THE STORY (MYANMAR LANGUAGE, SHORT, CLEAN) ================================================================ Write a Chinese Zen/Wuxia folktale based on the given Title. Rules: - LANGUAGE: The entire script must be written in clean, natural Myanmar (Burmese) language ONLY. Do not output any Chinese characters or Chinese-language text anywhere in the script — character names and place names should be written in Myanmar transliteration (e.g. လီ၊ ဝမ်၊ ကျန်း), never in Chinese script. - WRITING QUALITY: The prose must be smooth, grammatically correct, natural-flowing Myanmar — equivalent in polish to Gemini's writing quality. Before finalizing, proofread the entire script and eliminate any spelling errors, broken words, nonsensical phrases, or garbled/malformed sentences. Every sentence must make clear logical sense. - LENGTH: This script will become a 1–2 minute voice narration played at 1.5x speed, so keep it SHORT — approximately 350–450 Myanmar words total. Do not write a long story. Prioritize a tight, complete narrative arc (setup → event → Chan teaching/twist → closing insight) within this short length. - Cultural authenticity: the story must be strictly and purely Chinese in setting, character names, terminology, geography, clothing, architecture — no Japanese, Korean, Western, or any other country's elements anywhere. Use "Chan" 禪 instead of "Zen" in concept (described naturally in Myanmar, without needing the English/Chinese term itself). - STYLE: continuous narrative storytelling flow, no headers, no bullet points, no markdown symbols — plain flowing paragraph text only, written to sound natural and expressive when read aloud (TTS-ready) - TONE: warm, conversational storytelling voice, not academic or stiff - Weave the moral/teaching naturally into the ending, without a separate "Moral:" label Save this as: script.txt ================================================================ STEP 2 — BUILD A CHARACTER REFERENCE SHEET + LOCK ART STYLE (MANDATORY, DO THIS FIRST) ================================================================ Before generating any images, identify every recurring character in the story (e.g. the monk, the bandit) and lock down a fixed, detailed description for each one. This description must be written ONCE and then copied verbatim into every single image prompt that character appears in — never rephrase or vary it between scenes. For each recurring character, define and lock: - Exact face shape, age, skin tone, eye shape/color, eyebrows - Exact hairstyle, hair color, hair length, facial hair (if any) - Exact clothing: garment type, color, pattern, layering, accessories — same outfit throughout unless the story explicitly changes it - Build/height/posture - Any unique identifying marks (scar, jewelry, staff, weapon, etc.) LOCKED ART STYLE (use this exact style block, identically, in all 10 image prompts): A traditional Chinese gongbi (工笔) fine-line court painting in the style of Qing dynasty (清代) 仕女画 (shinu hua / court ladies painting), rendered on aged silk or linen scroll texture with a warm beige/tan background. Style requirements: - Meticulous fine ink outlines (细线勾勒) with flat, layered mineral-pigment coloring - Elegant elongated figures with pale porcelain-like skin, delicate facial features, small red lips, thin arched eyebrows - Elaborate traditional Hanfu/Qing-era robes with intricate embroidered patterns, flowing sleeves, sashes tied at the waist - Ornate hairstyles piled high with hairpins, combs, and floral or beaded hair ornaments - Muted, harmonious color palette: soft greens, dusty blues, terracotta reds, ochre yellows, ivory whites - Wooden furniture with red-brown lacquer finish, subtle inlaid details - No harsh shadows or Western-style perspective — flat decorative composition typical of classical Chinese painting - Aged parchment/silk texture background, slightly muted tones as if centuries old - Fine detail in textile patterns, jewelry, and hair accessories - Overall mood: refined, serene, domestic elegance of imperial court life Aspect ratio: portrait (vertical scroll format) If the generation tool supports reference images, character reference upload, or consistent seed/character-lock features, USE the same reference image or seed for that character across all 10 generations instead of relying on text description alone. ================================================================ STEP 3 — GENERATE 10 SCENE IMAGES ================================================================ Break the story into 10 key scenes, in story order. For each scene, build the image prompt by combining: 1. The LOCKED ART STYLE block from Step 2 (copied identically every time, word-for-word) 2. The FIXED character description block(s) from Step 2 for every character appearing in that scene (copied identically every time, word-for-word) 3. Scene-specific details only: setting, pose, action, mood for that particular moment (keep lighting flat/even, no harsh shadows, consistent with the gongbi style) Then actually GENERATE the image (not just the prompt text) for each of the 10 scenes: - Aspect ratio: 9:16 (vertical/portrait, scroll format) - Before finalizing each image, visually cross-check it against previously generated images of the same character — face, hair, and outfit must match. Regenerate if the character's appearance drifts from the locked reference. - Each image must visually match its corresponding moment in script.txt Save the 10 generated images as: image_01.png, image_02.png, ... image_10.png ================================================================ STEP 4 — GENERATE THE VOICE NARRATION (MP3) ================================================================ Convert script.txt into a spoken audio narration. Rules: - MODEL: Use Google's AI Studio / Google TTS model specifically for voice generation (a Google-model voice only — do not use ElevenLabs or any other non-Google TTS provider). - VOICE STYLE: Expressive, warm storytelling narration — natural intonation, good pacing and emotional tone suited to a Zen folktale (not flat or robotic). - PLAYBACK SPEED: 1.2x speed. - FINAL DURATION: the resulting audio must be approximately 1–2 minutes long at 1.2x speed. Save as: narrative.mp3 ================================================================ STEP 5 — GENERATE IMAGE-TO-VIDEO PROMPTS (SINGLE FILE, MYANMAR HEADINGS ONLY, ENGLISH PROMPT TEXT, FLOATING ANIMATION STYLE) ================================================================ For each of the 10 scenes/images, write one image-to-video generation prompt. LANGUAGE RULE: - Scene headings only must be in Myanmar (e.g. "မြင်ကွင်း ၀၁ (image_01.png)") - The actual prompt content itself (everything below each heading — style, motion, camera, audio description) must be written entirely in ENGLISH. Do not translate or write the prompt body in Myanmar — video generation models read English prompts more reliably. MOTION STYLE — because the source images are flat gongbi/court paintings with no Western perspective, animate them as a "living painting" rather than a cinematic 3D scene: - Camera must stay static or move with only extremely subtle, almost imperceptible drift — never a strong push-in, pan, or parallax move that would break the flat painterly illusion - Animate only surface-level elements: silk/hanfu fabric and sleeves swaying gently, hair ornaments and hairpins swinging softly, loose hair strands drifting, sashes/ribbons floating, incense smoke or mist curling slowly, flower petals or leaves floating gently through the frame - Motion strength: very low — slow, dreamlike, weightless floating quality throughout - Character's pose and framing should stay as close to the original painting composition as possible IMPORTANT — output format: combine ALL 10 prompts into ONE SINGLE text file (not 10 separate files). Structure the file like this, with each entry clearly separated: မြင်ကွင်း ၀၁ (image_01.png) [full image-to-video prompt text for scene 1 — written in English] မြင်ကွင်း ၀၂ (image_02.png) [full image-to-video prompt text for scene 2 — written in English] ... continue through... မြင်ကွင်း ၁၀ (image_10.png) [full image-to-video prompt text for scene 10 — written in English] Rules for each individual prompt: - STRICT WORD LIMIT: must NOT exceed 600 words. Count words before finalizing each prompt and trim if over. - Written entirely in English (only the heading line above it is in Myanmar) - Clearly reference which image file it applies to (as shown above) - Follow the MOTION STYLE rules above exactly (floating/living-painting animation, static or near-static camera) - Reuse the same locked character appearance details so any character motion stays visually consistent with the source image - End each prompt with an Audio line: no text, no captions, no music, no dialogue, no vocals — followed by a description of natural ambient sound only (wind, distant birds, faint fabric rustle, soft chimes, etc., matching the scene) Save as: img_to_video_prompts.txt ================================================================ STEP 6 — PACKAGE AND DELIVER ================================================================ Combine all output files into a single ZIP archive named: [Title]_zen_package.zip The ZIP must contain exactly these 4 items: - script.txt (Myanmar language, clean, proofread) - narrative.mp3 (Google TTS model voice, 1–2 min final duration, 1.2x speed) - image_01.png through image_10.png (9:16, gongbi court painting style, character-consistent) - img_to_video_prompts.txt (single file, all 10 prompts, Myanmar scene headings only, English prompt text, floating/living-painting motion style, referencing each image) Deliver only the final ZIP file as output. Do not output the script, prompts, or file contents separately in chat.
Tip — [TITLE] ထည့်ရုံဖြင့် script.txt + narrative.mp3 + ပုံ ၁၀ ပုံ + video prompts အကုန် ထွက်ပါမယ်။