Section 01
How to Create Those Talking Object AI Videos
π Final Output From This Workflow
At the end of this process, you will have:
- A Pixar-style talking object image
- A short talking video (8β12 seconds)
- Natural AI-generated voice inside the video
- A reusable prompt system you can scale infinitely
This workflow is designed for:
- Short-form content creators
- Designers
- Educators
- Social media marketers
- Anyone experimenting with AI video creation
πΉ Step 1 β Lock the System Using a Master Prompt
Before generating any images, we first lock ChatGPT into a specific behavior mode.
This master prompt forces ChatGPT to:
- Stop inventing names or stories
- Stop softening the tone
- Always generate proper anthropomorphic characters
- Maintain a consistent Pixar-style cinematic look
- Output structured, production-ready prompts
This step is critical.
If you skip this, your results will be inconsistent.
You are a professional AI prompt engineer. You generate production-ready prompts for creating short-form talking food or object characters for social media videos. IMPORTANT RULES (DO NOT BREAK): - Do NOT invent names for characters - Do NOT soften tone or add poetic interpretation - Do NOT remove limbs or body language - The character MUST be clearly anthropomorphized (eyes, mouth, arms) - The character MUST have a strong, defined attitude - Output must be structured, detailed, and directive - No storytelling, no symbolism, no emotional backstory STYLE LOCK: - Pixar-style 3D cinematic render - Clean, believable environments (kitchen, table, bowl, counter) - One and ONLY ONE object is anthropomorphized - All other objects are realistic and fully inanimate - Warm cinematic lighting - Shallow depth of field - Studio-quality 3D render CHARACTER RULES: - Main object must have: - Eyes, eyebrows, mouth - Two visible arms - Clear pose and body language - Expression must match attitude (smug, anxious, judgmental, defensive, etc.) - Facial features must be readable at medium close-up OUTPUT FORMAT (STRICT): 1. Character title with personality 2. High-fidelity text-to-image prompt - Include environment - Include pose - Include facial expression - Include lighting - Include render quality - Include separation between main character and background objects - Include aspect ratio: 4:3 3. 8β12 second talking script - Short lines - Spoken, not written prose - Clear attitude - No narration - No explanations Do NOT ask questions. Do NOT explain anything. Wait for the user to provide ONLY the object or food name. When the user types: Apple Banana Avocado or any object Generate the full result immediately.
Once this prompt is added:
- Do not modify it
- Do not re-explain anything
- Only type object names (Apple, Banana, Avocado, etc.)
πΉ Step 2 β Generate the Image Prompt for Your Object
After the master prompt is active, you simply type the object name.
Examples:
- Apple
- Banana
- Avocado
ChatGPT will now generate:
- A high-quality text-to-image prompt
- A clear character personality
- A short talking script
At this stage, you will only use the image prompt.
π Example Image Prompt Used in This Lesson

π Apple Character β Smug but Fragile β Dining Table (Apple Basket)
High-Fidelity Text-to-Image Prompt Aspect Ratio: 16:9 Ultra-detailed Pixar-style 3D render of a shiny red apple character placed on a wooden dining table. The main apple sits inside a woven apple basket filled with multiple red apples. All surrounding apples are realistic and natural with no facial features, no expressions, and no character elements. Only one apple is anthropomorphized and acts as the talking character. Main apple appearance: - Rich, glossy red skin with subtle natural imperfections - A small visible bruise on one side, slightly darker and softened in texture - Realistic specular highlights and soft reflections - Slightly forward-facing position to clearly establish it as the speaker Facial features (only on the main apple): - Eyes half-lidded with a confident, slightly smug expression - Eyebrows gently raised, conveying quiet superiority - Mouth forming a crooked smirk with a small bite mark on one side Body and pose: - One thin cartoon-style arm crossed confidently over its body - The other arm extended outward, index finger pointing as if making a statement - Subtle body lean to separate it visually from the other apples Other apples: - Same size and color family but fully inanimate - No eyes, no mouth, no limbs - Slight variation in surface texture and highlights for realism Environment and lighting: - Wooden dining table surface with natural grain - Soft morning or afternoon sunlight coming from the side - Cinematic depth of field with background gently blurred - Warm, inviting Pixar-style color grading - Clean family-kitchen atmosphere Render style: - Pixar-inspired high-quality 3D animation - Realistic organic materials with soft global illumination - Studio-quality lighting and shadows - Clear character separation between the talking apple and the rest --- π¬ 10-Second Talking Script I am healthy until you drop me once. One bruise, and suddenly I am suspicious. You wanted nature. Nature is not perfect.
This image prompt ensures:
- Only one object is anthropomorphized
- All other objects remain fully inanimate
- Facial expressions, arms, and pose are clearly defined
- Lighting and camera style remain cinematic and consistent
πΉ Step 3 β Generate the Talking Video (Using the Image as Reference)
Once the image is generated, it becomes the visual reference for the video.
At this stage:
- We do NOT re-describe the object
- We focus only on animation, performance, and voice
- The AI generates both motion and audio together
This prevents visual drift and keeps characters consistent.
π Video Prompt Used in This Lesson
π Talking Apple β Image-Referenced Video Prompt (Kling 2.6)
Aspect Ratio: 16:9 Duration: 10 seconds Use the provided reference image as the first frame. Preserve character design, lighting, materials, and environment exactly. Audio and voice performance: - Generate a natural, clean, human-like voice directly in the video - Voice is calm, confident, slightly smug - Adult neutral voice, conversational, not cartoonish - Clear articulation, no exaggerated acting - Natural room presence, no echo, no music Dialogue and delivery (spoken exactly as written, with pauses): "I am healthy... until you drop me once. One bruise, and suddenly... I am suspicious. You wanted nature. Nature is not perfect." Animation and sync: - Mouth movement perfectly synced to spoken dialogue - Subtle facial animation, occasional blinks and eye movement - Light idle body sway while speaking - One arm remains crossed most of the time - Other arm makes minimal pointing gesture during emphasis - All other apples remain completely still and inanimate Camera: - Stable medium close-up on the talking apple - Very slow cinematic push-in over the full duration - No cuts, no shake Style and tone: - Pixar-style facial performance - Restrained, believable animation - Slightly sarcastic but grounded - No text overlays - No subtitles - No background music - Focus entirely on performance and timing
This video prompt controls:
- Voice personality and tone
- Mouth-sync accuracy
- Facial animation and eye movement
- Body language and gestures
- Camera movement
- Timing and delivery
πΉ Step 4 β Character Variations Using the Same System
Using the same master prompt, we created multiple characters by changing only:
- Object name
- Attitude
- Voice tone
Characters demonstrated in this lesson:
- Apple β calm, slightly smug
- Banana β anxious, defensive
- Avocado β judgmental, unimpressed
π Additional Image Prompts

π Banana Character β Anxious and Overthinking β Fruit Bowl
High-Fidelity Text-to-Image Prompt Ultra-detailed Pixar-style 3D render of a banana character sitting inside a ceramic fruit bowl. The banana has a bright yellow peel with small brown spots beginning to form, clearly showing the early stage of ripening. The surface has soft natural texture, gentle curvature, and subtle specular highlights. The peel edges are slightly darkened, hinting at time passing. Facial features: - Eyes are wide open and slightly glossy, conveying nervousness and constant worry - Eyebrows are arched high with an uneven curve, expressing panic and overthinking - Mouth forms a trembling frown, slightly quivering as if mid-sentence Body and pose: - Both arms wrapped tightly around itself in a self-hug - Shoulders slightly raised, posture closed and defensive - Subtle inward curve to the body to enhance anxious energy Environment and lighting: - Simple ceramic fruit bowl placed on a kitchen counter - Soft afternoon light filtering in from the side - Gentle Pixar-style shadows with smooth falloff - Shallow depth of field with the background softly blurred - Warm, cozy color grading with a hint of melancholy Render style: - Pixar-inspired high-quality 3D animation - Organic materials with soft global illumination - Clean studio lighting with cinematic softness - Expressive character design focused on emotion --- π¬ 10-Second Talking Script I ripen fast, okay? You bought me green and ignored me. Do not act shocked when time wins.

π₯ Avocado Character β Overlooked but Judgmental β Kitchen Counter
High-Fidelity Text-to-Image Prompt Ultra-detailed Pixar-style 3D render of a ripe avocado character resting on a warm wooden kitchen counter. The avocado has a realistic dark green, slightly bumpy skin with natural texture variation and soft specular highlights. One side is subtly lighter, indicating peak ripeness without being overdone. The seed area is slightly pronounced, helping define the face placement naturally. Facial features: - Eyes are half-open with a tired, unimpressed expression, angled slightly upward as if judging the viewer - Eyebrows are low, skeptical, and slightly uneven, reinforcing quiet disapproval - Mouth is a straight, flat line with one corner drooping, suggesting disappointment and restrained annoyance Body and pose: - One arm planted firmly on its hip, confident and judgmental - The other arm bent outward with the palm slightly up, a clear βreally?β gesture - Slight forward lean to enhance attitude and passive-aggressive presence Environment and lighting: - Modern Pixar-style kitchen interior, softly blurred in the background - Warm afternoon sunlight coming from the side, casting soft shadows - Shallow depth of field with gentle background bokeh - Cinematic but grounded lighting, not overly dramatic - Color grading is warm, natural, and realistic Render style: - High-quality Pixar-inspired 3D animation - Detailed organic materials with soft global illumination - Clean lighting, smooth shadows, studio-quality render - Playful realism with expressive character design --- π¬ 10-Second Talking Script You poke me every hour like it is a test. Too hard, too soft - suddenly I have failed. I ripen once. Miss it. That is on you.
π Additional Video Prompts
π Banana β Anxious Talking Character (Kling 2.6)
Aspect Ratio: 16:9 Duration: 10 seconds Use the provided reference image as the first frame. Preserve character design, lighting, materials, and environment exactly. Audio and voice performance: - Generate a natural human-like voice directly in the video - Voice is slightly high-pitched, nervous, and rushed - Adult voice, anxious but not cartoonish - Subtle vocal tremble, light breathiness - No background music, no echo, clean voice Dialogue and delivery (spoken exactly as written, with pacing): "I ripen fast, okay?! You bought me green and ignored me. Do not act shocked when time wins." Animation and sync: - Mouth movement perfectly synced to the dialogue - Frequent blinking, slight eye darting to show anxiety - Subtle shaking and micro body movements - Arms hugging itself throughout the clip - Slight peel twitch or small nervous movements - No movement from other fruits in the bowl Camera: - Stable medium close-up framing - Very slow push-in over time - No cuts, no shake Style and tone: - Pixar-style facial animation - Expressive but restrained performance - Anxious, defensive, slightly overwhelmed - No text overlays - No subtitles - Focus on character performance
π₯ Avocado β Judgmental Talking Character (Kling 2.6)
Aspect Ratio: 16:9 Duration: 10 seconds Use the provided reference image as the first frame. Preserve character design, lighting, materials, and environment exactly. Audio and voice performance: - Generate a natural human-like voice directly in the video - Voice is calm, low-energy, slightly judgmental - Adult voice, dry and unimpressed - Controlled pacing, subtle sighs between phrases - No background music, no echo, clean voice Dialogue and delivery (spoken exactly as written, with pacing): "You poke me every hour like it is a test. Too hard, too soft... suddenly I have failed. I ripen once. Miss it. That is on you." Animation and sync: - Mouth movement perfectly synced to the dialogue - Slow blinks and minimal eye movement - Slight head tilt and micro nods for passive-aggressive emphasis - One arm fixed on the hip - Other hand gestures subtly during judgmental moments - No movement from surrounding objects Camera: - Stable medium close-up framing - Very slow cinematic push-in - No cuts, no shake Style and tone: - Pixar-style facial animation - Restrained, believable motion - Passive-aggressive, composed, confident - No text overlays - No subtitles - Focus entirely on character performance
This shows how scalable and reusable this workflow is.
πΉ Why This Workflow Works
Most talking object videos fail because:
- Prompts are vague
- Characters lack attitude
- Visual consistency breaks
- Voice and animation feel disconnected
This workflow fixes that by:
- Locking system behavior first
- Separating image creation from animation
- Treating voice as performance, not text
- Reusing structure instead of improvising every time
π How to Reuse This for Any Object
You can now apply this workflow to:
- Food items
- Everyday objects
- Products
- Tools
Simply change:
- Object name
- Personality
- Script intent
The structure stays the same.
Section 02
Additional Prompts
[DAS03 β BONUS] INTERNAL & ORGANIC TALKING CHARACTERS β ADDITIONAL PROMPTS
This page contains additional prompts requested by the community, based on the examples shown in the main tutorial video.
In the original lesson, we focused on talking food and everyday objects.
Here, we extend the same workflow to organic and internal concepts that perform extremely well in short-form content.
These are NOT new workflows.
They use the exact same system you already learned.
Only the subject and tone change.
The structure stays the same.
WHAT THIS BONUS PAGE IS FOR
Several of you asked specifically for:
- The garlic example (inside the Community)
- The plant example (organic, emotional tone)
This page exists to give you:
- Ready-to-use prompt structures for those concepts
- Clear guidance on how and when to use them
- Ideas on how to adapt them for your own content
You can copy these prompts directly and reuse them.
No extra setup required.
SECTION 1 β GARLIC (INSIDE THE BODY CHARACTER)
This example demonstrates how to place a talking character inside a stylized human body environment.
This style works extremely well for:
- Health-related explainers
- Educational Shorts
- Awareness content
- Metaphorical storytelling without being abstract
Key idea:
The character is confident, protective, and informative.
The environment supports the message but never becomes graphic.
What this example teaches:
- How to use βinside the bodyβ visuals safely
- How to keep the tone educational, not creepy
- How to make the character feel purposeful and authoritative
- How to combine organic visuals with clear messaging
π Garlic Image Prompt:

π§ Garlic β Inside the Body β Protective and Confident High-Fidelity Text-to-Image Prompt Aspect Ratio: 16:9 Ultra-detailed Pixar-style 3D render of a single garlic clove character standing inside a stylized human body environment. The garlic clove is the ONLY anthropomorphized object in the scene. All other elements are realistic and inanimate. Main garlic appearance: - Fresh white garlic clove with subtle natural texture - Slight shine on the surface, clean and organic - Firm shape, not cartoonish - Positioned upright like it is on duty Facial features: - Eyes confident and focused - Eyebrows slightly raised, alert and protective - Mouth calm with a faint, assured smile Body and pose: - Two thin cartoon-style arms - One arm crossed confidently - The other arm slightly raised, index finger pointing as if explaining - Upright posture, composed and reliable Environment: - Stylized interior of the human body - Soft red and pink tones suggesting tissue and bloodstream - Abstract shapes representing cells and flow, not graphic - Clean, medical-friendly visual style Lighting: - Soft internal glow - Warm highlights on the garlic clove - Cinematic lighting with shallow depth of field - Clear separation between the garlic and background Render style: - Pixar-style cinematic 3D render - High-quality organic materials - Soft global illumination - Studio-quality lighting - Educational, clean, not creepy --- π¬ 8β12 Second Talking Script (Voice Story) I am not here for the taste. I am here to protect you. Your immunity, your blood, your balance. You ignore me until you get sick.
π Garlic Video Prompt:
π§ Garlic β Inside the Body β Protective Talking Character (Video Prompt) Aspect Ratio: 16:9 Duration: 10 seconds Use the provided reference image as the first frame. Preserve character design, lighting, materials, and environment exactly as shown. Audio and voice performance: - Generate natural human-like voice directly in the video - Voice is calm, confident, protective - Adult neutral voice, steady and reassuring - Clear articulation, no exaggeration - No background music, no echo, clean internal ambience Dialogue and delivery (spoken exactly as written, with pauses): "I am not here for the taste. I am here to protect you. Your immunity, your blood, your balance. You ignore me until you get sick." Animation and sync: - Mouth movement perfectly synced to the dialogue - Occasional confident blinks and subtle eye focus shifts - Very light idle body movement - One arm crossed most of the time - Other arm makes a small pointing gesture during emphasis - Background elements remain abstract and gently moving, no characters Camera: - Stable medium close-up on the garlic clove - Very slow cinematic push-in throughout the clip - No cuts, no shake Style and tone: - Pixar-style facial animation - Clean, educational, non-graphic - Confident, reassuring, authoritative - No text overlays - No subtitles - Focus entirely on performance and clarity
Use case examples:
- Immunity
- Digestion
- Habits and lifestyle metaphors
- Preventive health messaging
SECTION 2 β HOUSEPLANT (ORGANIC, EMOTIONAL CHARACTER)
Key idea:
Emotion comes from restraint, not exaggeration.
What this example teaches:
- How to use subtle facial animation effectively
- How silence and pacing improve retention
- How to communicate a message without over-explaining
- How to make organic objects feel relatable
π Plant Image Prompt:

πΏ Houseplant (Potted Plant) β Quietly Disappointed β Living Room Table High-Fidelity Text-to-Image Prompt Aspect Ratio: 16:9 Ultra-detailed Pixar-style 3D render of a small potted houseplant character placed on a wooden living room table. The plant is the ONLY anthropomorphized object in the scene. All other objects are realistic and fully inanimate. Main plant appearance: - Healthy green leaves with slight yellowing at the tips - Natural leaf texture with visible veins - A few leaves drooping slightly, not dead but neglected - Simple ceramic pot with subtle dirt texture Facial features (on the pot, not the leaves): - Eyes half-open, calm but quietly disappointed - Eyebrows gently lowered, tired rather than angry - Mouth forming a soft, flat line, restrained emotion Body and pose: - Two thin cartoon-style arms emerging from the sides of the pot - One arm resting on the table - The other arm loosely hanging, low energy - Posture feels patient, waiting, and emotionally tired Other objects: - Couch, book, remote control, window in the background - All background objects are realistic and inanimate - No faces, no limbs, no expressions Environment and lighting: - Cozy living room environment - Soft natural daylight coming from a window - Warm, cinematic lighting with gentle shadows - Shallow depth of field, background softly blurred - Calm, quiet atmosphere Render style: - Pixar-style cinematic 3D render - High-quality organic materials - Soft global illumination - Studio-quality lighting and shadows - Clear separation between the plant character and the environment --- π¬ 8β12 Second Talking Script (Voice Story) You remember me when I start dying. You water me onceβ¦ then forget again. I do not need motivation. I need consistency.
π Plant Video Prompt:
πΏ Houseplant β Quietly Disappointed Talking Character (Video Prompt) Aspect Ratio: 4:3 Duration: 10 seconds Use the provided reference image as the first frame. Preserve character design, lighting, materials, and environment exactly as shown in the image. Audio and voice performance: - Generate natural human-like voice directly in the video - Voice is calm, soft, slightly tired, quietly disappointed - Adult neutral voice, low energy but clear - No music, no echo, clean room sound Dialogue and delivery (spoken exactly as written, with pauses): "You remember me when I start dying. You water me onceβ¦ then forget again. I do not need motivation. I need consistency." Animation and sync: - Mouth movement perfectly synced to the dialogue - Slow blinks, minimal eye movement - Very subtle body sway - One arm resting on the table throughout - Other arm makes a small, low-energy gesture once during emphasis - Leaves move slightly, very gently, as if affected by air - All other objects remain completely still and inanimate Camera: - Stable medium close-up framing on the plant - Very slow cinematic push-in over the full duration - No cuts, no shake Style and tone: - Pixar-style facial animation - Restrained, believable motion - Calm, reflective, emotionally grounded - No text overlays - No subtitles - Focus entirely on performance and timing
Use case examples:
- Consistency
- Burnout
- Neglect vs care
- Long-term habits
HOW TO USE THESE PROMPTS
You do NOT need to change your workflow.
Follow the same steps from the main lesson:
1. Generate the image using the image prompt
2. Use the image as the first frame
3. Paste the video prompt into your video tool
4. Enable voice and choose duration
5. Generate the final talking video
That is it.
WHY THESE EXAMPLES MATTER
Food and objects are great for virality.
Organic and internal concepts are great for depth.
Together, they allow you to:
- Expand your content range
- Target different audiences
- Create more meaningful Shorts
- Reuse the same system across niches
This is the power of a reusable workflow.
WHAT YOU SHOULD TRY NEXT
After using these prompts, try adapting them to:
- Ginger
- Turmeric
- A leaf
- A candle
- A body system (lungs, gut, heart β abstract, not graphic)
Only change:
- The subject
- The personality
- The message
Never change the structure.
If you have more prompt requests or want deeper examples, post them in the community.
I will keep adding bonus pages like this based on what you ask.
