提示词

{ "$schema": "farmix-veo-omni-image-to-video-prompt-v2", "prompt_metadata": { "target_model": "Google Veo 3 / Omni (image-to-video). Also compatible with Kling (Elements / start-frame) and Runway (Gen-4 References).", "task": "Produce a FINISHED, ready-to-post 10-second Instagram Reel for Farmix. No post-production, no compositing, no editing by the user.", "generation_mode": "IMAGE-TO-VIDEO. The uploaded images are fed as real inputs and animated. The model must NOT redraw or translate any text or logo — it animates the actual input pixels.", "version": "2.0-Farmix-ReadyToPost", "why_image_to_video": "Text-to-video would garble Arabic UI text and the logo. By driving each app/logo shot from the real uploaded image, the exact Arabic screens and the Farmix logo are preserved perfectly with zero manual work.", "assembly_note": "Veo generates ~8s max per clip, so the reel is 4 short clips placed back-to-back (simple concatenation, no editing skills needed). To get the full 10s in a SINGLE generation with no joining at all, use Kling (supports 10s clips)." }, "reference_images_as_inputs": { "home.jpg": { "role": "REAL app home screen (Arabic RTL, cream UI). Used as the input frame for shot_02.", "rule": "Preserve exactly. Do NOT regenerate, translate, restyle, or add/remove any text, icon, or element. Animate only." }, "login.jpg": { "role": "REAL login / onboarding screen (Arabic RTL, +963, Farmix logo). Used as the input frame for shot_03.", "rule": "Preserve exactly. Do NOT regenerate, translate, restyle, or alter any text/logo. Animate only." }, "logo_sheet.png": { "role": "Farmix logo system. Use the horizontal green lockup as the input for the shot_04 end-card reveal.", "rule": "Preserve exact logo shape, colors, and the word 'Farmix'. Do NOT redraw or distort. Animate only." } }, "global_specification": { "aspect_ratio": "9:16", "dimensions": { "width": 1080, "height": 1920 }, "fps": 30, "total_duration_seconds": 10, "output": "H.264, high bitrate, ready for Instagram Reels.", "status": "PRE-LAUNCH. Keep the tone aspirational and brand-forward; do not over-claim a live service." }, "on_screen_text_policy": { "model_generates_new_text": false, "rule": "The ONLY visible text/logo in the video comes from the input images (the real app screens and the Farmix wordmark). The model must not invent, draw, caption, or subtitle any additional Arabic or Latin text. All extra messaging is delivered by voiceover only." }, "art_direction": { "aesthetic": "Warm cinematic editorial. Clean, premium, restrained. Natural golden-hour light, shallow depth of field.", "camera_language": "Slow, confident motion only — gentle push-ins, subtle parallax, soft orbits. No shakes or whip-zooms.", "grade": "Warm golden editorial; softly lifted shadows with a faint forest-green (#004630) tint; protected cream highlights; light natural grain.", "forbidden": ["glassmorphism", "cheap/neon gradients", "over-saturation", "heavy flares", "warped hands/fingers", "any model-drawn text", "watermarks", "cold or dark mood"] }, "shot_list": [ { "id": "shot_01_hook", "duration_seconds": 2.5, "mode": "text-to-video (no input image needed; no on-screen text)", "setting": "A stunning private Syrian farm villa with a turquoise pool, greenery and warm stone at golden hour.", "camera": "Slow cinematic push-in toward the pool, shallow depth of field, gentle drift.", "action": "Sunlight shimmering on calm water, soft breeze in the trees — an irresistible summer-escape feeling.", "voiceover": { "line": "هناك، حيث تهدأ الحياة قليلاً...", "tone": "hushed, cinematic, unhurried; elegant Modern Standard Arabic" }, "sfx": "Soft water shimmer, warm ambient birds/breeze." }, { "id": "shot_02_app_home", "duration_seconds": 3.0, "mode": "image-to-video", "input_image": "home.jpg", "preserve_input_exactly": true, "do_not_regenerate_or_translate_text": true, "presentation": "Show the real home screen as a living app screen filling a clean vertical frame (screen-recording style) OR seated inside a simple modern phone that frames it — but the SCREEN CONTENT is exactly home.jpg, untouched.", "animation": "Bring the static screen to life: a slow, natural upward scroll revealing the listings, a gentle pointer/thumb tap on the highlighted villa card, subtle depth/parallax on cards. Motion must look like real, smooth app usage. No loading states.", "camera": "Very slight push-in on the screen; keep it sharp, upright, fully readable.", "voiceover": { "line": "فارمكس يفتح لك أبواب أرقى المزارع والشاليهات في سوريا؛ تتصفّح، تختار، وتحجز بكل سهولة.", "tone": "refined, confident, clear MSA; gentle emphasis on أرقى" }, "sfx": "Soft UI tick on the tap, warm bed." }, { "id": "shot_03_login_join", "duration_seconds": 2.0, "mode": "image-to-video", "input_image": "login.jpg", "preserve_input_exactly": true, "do_not_regenerate_or_translate_text": true, "presentation": "Show the real login / onboarding screen exactly as provided (Arabic RTL, +963, Farmix logo intact).", "animation": "Gentle life: soft entrance of the phone-number field and buttons with subtle spring, a light glow pulse on the primary button, faint parallax on the background image grid. Nothing regenerated — animate the real pixels only.", "camera": "Slow, steady slight rise.", "voiceover": { "line": "خطواتٌ بسيطة، وتصبح وجهتك جاهزة. وإن كنت صاحب مزرعة، فمكانك في الصدارة.", "tone": "warm yet polished MSA; slight lift on الصدارة" }, "sfx": "Soft confirm chime, warm bed." }, { "id": "shot_04_logo_end_card", "duration_seconds": 2.5, "mode": "image-to-video", "input_image": "logo_sheet.png (horizontal green lockup)", "preserve_input_exactly": true, "do_not_regenerate_or_translate_text": true, "presentation": "Clean warm cream (#f7f5e9) background with soft golden light; the real Farmix logo centered, preserved exactly.", "animation": "Elegant logo reveal: the logo settles in with a soft scale/settle and a gentle light bloom passing across it. Calm, premium, suspenseful. No extra text drawn by the model.", "camera": "Very slow push-in, then hold on the logo.", "voiceover": { "line": "فارمكس... حيث تبدأ إجازتك.", "tone": "calm, premium resolve; pause before حيث تبدأ إجازتك so the brand name lands" }, "sfx": "Soft rising shimmer / warm resolve chord." } ], "audio_design": { "music": { "style": "Warm, modern, light-cinematic bed; acoustic + soft electronic; hopeful summer feel; ~100 BPM.", "mix": "Ducked under voiceover; lifts on the end card; cuts land on the beat (~2.5s, ~5.5s, ~7.5s)." }, "voiceover": { "language": "Elegant, refined Modern Standard Arabic. Cinematic luxury-brand tone — warm, calm, and premium. Not colloquial.", "note": "If the model's spoken Arabic is not clean, generate the clips muted and add the Levantine voiceover + music track over the finished cut — this is the only case where a light audio pass is needed; the visuals still require no editing." } }, "negative_prompt": [ "model-generated or redrawn text of any kind", "translated, restyled, or altered app UI text", "garbled / gibberish Arabic characters", "distorted or redrawn Farmix logo", "extra or warped fingers, distorted hands", "loading spinners, error states, empty/blank screens", "glassmorphism, cheap gradients, over-saturation", "watermarks, timestamps, added captions or subtitles", "shaky camera, whip-zooms, motion smear", "cold, dark, or cluttered ornamental mood" ], "how_to_run": { "veo_omni": "Run each app/logo shot in IMAGE-TO-VIDEO: upload the mapped image as the input frame, paste that shot's animation + camera + voiceover text as the prompt. Generate shot_01 as text-to-video. Place the 4 clips back-to-back (~10s). Keep audio if clean.", "kling": "Use start-frame image-to-video (or Elements) with the same images; Kling can render up to 10s, so you may produce it as fewer/one clip and skip joining.", "runway": "Use Gen-4 image-to-video / References with the mapped images as the driving frames.", "golden_rule": "Whenever a real screen or the logo must appear, it MUST be the input image, never described in words — that is what keeps everything ready-to-post with perfect Arabic." }, "delivery_assertions": [ "Final: 10.0s, 9:16, 1080x1920, ready to post.", "All visible screens and the logo are the real uploaded images, unaltered.", "No text or logo is generated by the model; extra messaging is voiceover-only.", "Tone is premium and brand-forward; no over-claiming a live service.", "No manual compositing required by the user." ] }

生成结果

Gemini Omni Video

参数

参考图片