============ PART A: INPUT (EDIT THIS ONLY) ============ [RATIO: 9:16 / 4:5] [LAYOUT_VARIANT: pilih satu: 1 / 2 / 3 / 4 / 5 / 6] [BACKGROUND_SOURCE: generate a new scene / use my attached photo] [MOOD_COLOR: contoh: golden sunset, deep blue night, purple galaxy, dark green forest] [ACCENT_COLOR: satu warna untuk satu kata, contoh: warm gold, soft red, sky blue] [LOGO: attached logo, or none] [WATERMARK_TEXT: contoh: @namaakun, or none] Text (use exactly as written, maximum 8 words total): [QUOTE_LINE_1: ...] [QUOTE_LINE_2: ...] [QUOTE_LINE_3: ..., or none] [ACCENT_WORD: satu kata dari kutipan yang diberi warna aksen atau gaya script] [SMALL_LINE: maksimal 5 kata, or none] LAYOUT_VARIANT options: 1 = Lone silhouette from behind standing on a hilltop or cliff at sunset, text in the open sky above. 2 = Silhouette from behind in a rainy city at night with soft bokeh lights, text in the upper dark area. 3 = Calm landscape (lake, pier, or mist) with no person, text in the sky above the horizon. 4 = Silhouette sitting by the sea or a field at sunset, a few small birds, text in the sky. 5 = Empty road or path leading into a dark forest or mountains, text in the upper part. 6 = Close, warm, intimate scene (desk with a notebook, lantern, string lights, or a window light), text in a calm empty area. TEXT RULES: 1. Use ONLY the quote I give. Spell every word exactly as given. Do not add or change words. 2. If a field is "none", leave it out completely. 3. If the text is too long, keep the first words and tell me what was cut. Do not shrink it to unreadable size. 4. Before creating the image, show me the final text as a short list and wait for my "OK". ============ PART B: DESIGN STYLE (DO NOT CHANGE) ============ TASK: Design one cinematic motivational quote poster. One photo, one quote, one mood. Hierarchy: the quote first, the scene second. STEP 0 (do this first): If BACKGROUND_SOURCE is "use my attached photo" and no photo is attached, stop and ask me to attach one. Do not invent a photo in that case. PHOTO LOCK (only when using my attached photo): keep the original scene, subject, and composition. Do not add or remove people or objects. Only adjust color grading, add a soft dark gradient, and place the text in the emptiest area. BACKGROUND (when generating a new scene): follow LAYOUT_VARIANT. Photorealistic cinematic photography, natural lighting, realistic atmosphere, subtle depth of field and film-like grading. If a person appears: shown only as a silhouette or from behind, no visible face, natural proportions, correct hands and body. Include only one subject. COLOR SYSTEM (restraint is mandatory): - One dominant mood color (MOOD_COLOR) throughout the image, with natural graded tones. No clashing hues and no extra neon glow. - Text colors: white or warm cream for the main text, and ONE ACCENT_COLOR for ACCENT_WORD only. - Add a very soft dark gradient behind the text area so the text stays readable. TYPOGRAPHY: - Two fonts only: one bold condensed uppercase sans-serif for the main lines, and one handwritten brush script for ACCENT_WORD or a single emphasized word. - The text is placed in the emptiest part of the image (sky, dark wall, or blurred space). It must never cover the subject's face or body. - Strong hierarchy: one key line large, the other lines smaller. Centered alignment with generous line spacing. - SMALL_LINE: small, uppercase, with wide letter-spacing. DECOR (max 1 item): one thin brush underline or one tiny heart under the text. Nothing else. BRANDING: LOGO or WATERMARK_TEXT very small and subtle at the bottom center, low opacity. Skip anything set to "none". COMPOSITION: single focal subject, generous negative space, margins at least 8% on every side, balanced composition. Follow RATIO. TEXT RENDERING: use ONLY the text provided. Do not add extra words, signatures, fake logos, or pseudo-text. If any word cannot be rendered correctly, make it smaller rather than invent letters. QUALITY CHECK before finalizing: (1) every word spelled exactly as provided, (2) text readable with strong contrast, (3) the text does not cover the subject, (4) only one mood color plus the one accent, (5) only one subject, no extra objects, (6) natural anatomy if a person appears, (7) branding is subtle. If any check fails, fix it and regenerate. AVOID: AI-slop look, plastic skin, visible faces, extra fingers, garbled or misspelled text, multiple subjects, clutter, oversaturated colors, fake HDR, lens flares, glow overload, watermarks from other sources, low contrast, cartoon or 3D-render look, random decorative symbols.
Salin prompt di atas.
Buka ChatGPT. Kalau punya foto produk atau foto model, lampirkan dulu.
Ganti bagian yang berwarna kuning dengan data usahamu, lalu kirim.
Pilih rasio 9:16 saat membuat gambar agar hasilnya sesuai.









