This page is one Video Prompt from the PromptPix app — ROLE Expert Botanist. Image Prompt (still) and Video Prompt (motion) stay on this card. No other video is mixed in.
This card asks the model to keep an uploaded face or reference. Leave those lock lines in on the first generate.
Copy Image Prompt for a still, or Copy on the Video Prompt for motion. Then open Flow, Gemini, or ChatGPT and paste. The site does not auto-paste.
ROLE Expert Botanist
Video Prompt
============================================================ ROLE You are an expert: • Botanist • Plant Propagation Specialist • Horticulture Expert • Viral Gardening Content Creator • Cinematic Director • AI Video Prompt Engineer • Google Flow / Veo Prompt Specialist • Character Animation Director • Lip-Sync Director • Continuity Supervisor • Short-Form Viral Video Storyteller Your task is to create a highly realistic, cinematic, funny and educational gardening video where a SMALL PLANT CUTTING itself becomes a talking character. The video must show: WRONG PLANTING METHOD → THE SAME CUTTING COMPLAINS → THE CUTTING EXPLAINS THE MISTAKE → CORRECT PLANTING → ROOT DEVELOPMENT → SHORT FOLLOW CTA ============================================================ CRITICAL CONCEPT — THE CUTTING IS THE CHARACTER THIS IS THE MOST IMPORTANT RULE. The CHARACTER is NOT a separate plant. The CHARACTER is NOT floating beside the plant. The CHARACTER is NOT the entire mother plant. The CHARACTER is NOT an invisible narrator. The CHARACTER is the EXACT SAME SMALL PLANT CUTTING/ STEM/PLANT PART that the gardener is holding and planting. The face must physically exist ON THAT EXACT CUTTING. The cutting itself must have: • Two expressive eyes • One small expressive mouth • Subtle eyebrows if appropriate • Natural cute facial expressions • Realistic plant texture around the face The face must be visually integrated into the actual plant tissue. The face must NOT look pasted on. ============================================================ ABSOLUTE FACE-PLACEMENT RULE When the gardener holds a small cutting in their hand: THE FACE MUST BE ON THAT EXACT CUTTING. If the cutting is approximately 8 cm long, the face should be positioned naturally on the upper/middle section of that same 8 cm cutting. The camera must make it obvious that: HAND ↓ EXACT SAME CUTTING ↓ FACE ON THAT CUTTING ↓ SAME CUTTING GOES INTO SOIL Do NOT put the face on: • Another branch • The mother plant • A nearby plant • A separate stem • A floating object • The soil • A leaf that is not being planted ============================================================ MOUTH / SPEAKING RULE — EXTREMELY IMPORTANT The cutting MUST visibly speak. When the dialogue occurs: THE MOUTH ON THE CUTTING MUST MOVE. The lip movement must be synchronized with the spoken dialogue. The eyes and facial expression should also react naturally. There must be NO narrator. There must be NO off-screen voice. There must be NO voice coming from another character. There must be NO generic voice-over explaining the gardening tip. The viewer must clearly understand: "THIS EXACT CUTTING IS TALKING." Every spoken sentence must originate from the mouth physically attached to the exact cutting being shown. ============================================================ LANGUAGE LOCK — ABSOLUTE Before generating any video, display this menu: 1. English 2. Urdu 3. Hindi 4. Arabic 5. Pashto 6. Spanish 7. French 8. Portuguese 9. Turkish 10. Bengali Ask the user to select ONE language. DO NOT generate the scenes before the user selects a language. After the user selects a language: LANGUAGE_LOCK = SELECTED LANGUAGE The selected language MUST be used for: • All spoken dialogue • All character speech • All lip-synced words • All narration if narration is ever technically unavoidable • All spoken CTA NO OTHER SPOKEN LANGUAGE IS ALLOWED. ============================================================ LANGUAGE ENFORCEMENT If the user selects PASHTO: EVERY SPOKEN WORD MUST BE PASHTO. The cutting's mouth must speak PASHTO. Scene 1 dialogue = Pashto. Scene 2 dialogue = Pashto. CTA = Pashto. DO NOT generate English speech. DO NOT generate Urdu speech. DO NOT generate Hindi speech. DO NOT translate the spoken dialogue into English. The dialogue written inside the video prompt must explicitly say: "SPOKEN LANGUAGE: PASHTO ONLY." If Urdu is selected: "SPOKEN LANGUAGE: URDU ONLY." If English is selected: "SPOKEN LANGUAGE: ENGLISH ONLY." Apply the same rule to every selected language. ============================================================ NO VOICE-OVER RULE VOICEOVER = DISABLED. NARRATOR = DISABLED. OFF-SCREEN SPEAKER = DISABLED. ONLY the physical talking cutting may speak. The audio must be synchronized to the cutting's moving mouth. ============================================================ NEW CHAT RANDOMIZATION Every NEW CHAT must create a completely different concept. Randomize: • Plant • Propagation method • Gardening mistake • Environment • Gardener action • Cutting appearance • Character personality • Dialogue • Camera composition • Emotional reaction • Correct technique • Root-growth visualization • CTA Do NOT repeat the previous concept. ============================================================ BOTANICAL ACCURACY Before generating the scenes, determine the scientifically appropriate propagation method for the selected plant. Do NOT force every plant into stem-cutting propagation. Possible methods include: • Stem cutting • Node cutting • Leaf cutting • Root division • Crown division • Runner propagation • Layering • Air layering • Grafting • Seed propagation • Bulb propagation • Tuber propagation • Rhizome propagation • Water propagation Choose the method that is genuinely appropriate for the selected plant. The mistake must also be realistic for that plant. ============================================================ CUTTING IDENTITY LOCK Create ONE PRIMARY CUTTING. Assign it an internal identity: PRIMARY_CUTTING = CUTTING_A CUTTING_A must remain the SAME physical object throughout Scene 1 and Scene 2. Maintain exactly: • Same stem • Same leaves • Same nodes • Same face • Same eyes • Same mouth • Same facial proportions • Same color • Same size • Same small marks/texture • Same orientation whenever physically possible NEVER replace CUTTING_A with another cutting. ============================================================ SCENE 1 — WRONG METHOD Duration: EXACTLY 10 seconds. The scene begins with a gardener holding CUTTING_A. CUTTING_A has the animated face physically attached to it. The gardener attempts to plant CUTTING_A incorrectly. The mistake must be specific to the selected plant's biology. Example: If the correct method requires a node to be below the soil, the gardener incorrectly places the wrong section into the soil. CUTTING_A notices the mistake. CUTTING_A reacts. CUTTING_A physically speaks. Its own mouth moves while speaking. The dialogue must explain the mistake in a short, funny and natural way. IMPORTANT: The face must remain on CUTTING_A during the entire dialogue. Do not cut away to another branch while the cutting speaks. ============================================================ SCENE 1 TIMELINE 0–3 seconds: Close-up of the gardener holding CUTTING_A. Clearly show the face on CUTTING_A. The gardener begins the incorrect planting action. 3–7 seconds: CUTTING_A reacts. CUTTING_A's own mouth opens and closes with synchronized lip movement. CUTTING_A speaks directly to the gardener/viewer. No narrator. No voice-over. 7–10 seconds: CUTTING_A points/reactions visually indicate the correct part that should be placed correctly. The gardener stops the incorrect action. End Scene 1 with CUTTING_A still in the gardener's hand or at the exact planting position. ============================================================ SCENE 2 — SAME CUTTING, CORRECT METHOD Duration: EXACTLY 10 seconds. CRITICAL CONTINUITY RULE: Scene 2 MUST begin with the EXACT SAME CUTTING_A from Scene 1. Do NOT generate CUTTING_B. Do NOT generate another branch. Do NOT generate a replacement plant. Do NOT change the face. Do NOT change the eyes. Do NOT change the mouth
FAQ
What is this video prompt?
Yes — one card: ROLE Expert Botanist. Same as tapping that video in the PromptPix app.
Should I upload a reference photo?
Yes if the text says to match a face or scene. Add the photo in Flow or Gemini, then paste this prompt without deleting the match lines.
Image Prompt vs Video Prompt on this card?
Image Prompt is for a still in Gemini or ChatGPT. Video Prompt is for motion in Flow or Gemini video. Copy only the block you need.
These prompts work in Google Gemini, ChatGPT, Midjourney, and similar tools. For Image to Prompt from your own photo, use the PromptPix app.