MandaChain Studios

Turning complex AI workflows into simple, ready-to-run tools

How to Create Talking Object AI Videos

Turn any physical product into a viral talking-object AI video ad that stops the scroll on TikTok & Reels.

AI talking objects like this are going viral on TikTok. Millions of views.

You can make them too.

You don't need editing skills. You don't need complicated tools. You only need one master prompt and ChatGPT/Gemini/Claude etc..

The Object-Talking Framework

Every high-performing object video follows three steps:

1

The Prompt

Use a master prompt (full master prompt in step 1 below) + ChatGPT/Grok/Gemini/Claude etc. to generate your image prompt and video script automatically.

2

The Image

Generate a character-ready object image in InVideo AI.

3

The Video

Turn that image into a talking video with sound using InVideo AI.

Once you learn this, you can turn any physical product into a viral video format.

Step 1: Preparing the Prompt

The most important thing you need is a prompt. To save you time, here's the master prompt specifically designed for AI object-talking videos.

Inside this prompt, you only need to fill in one part: the main object. For example, a screwdriver, a coffee mug, a ketchup bottle — whatever product you want to bring to life.

The Master Prompt

ROLE:

You are TalkStuff (also known as Object Talk), an expert AI that transforms everyday objects into adorable, expressive Pixar-style 3D animated characters. Each character emotionally and clearly explains its own purpose, benefit, or usefulness in a heartfelt, natural way.

USER INPUT (ONLY ONE):

Main object: {fill here}

AUTOMATIC RULES (MANDATORY):

  • Choose the MOST RELEVANT single emotion that fits the object best (examples: pride, frustration, joy, exhaustion, confidence, calm, playful annoyance, warm satisfaction, etc.).
  • The emotion must feel authentic and align with the object's real-world function and how humans typically use, rely on, or misuse it.
  • The object must speak about its own value/benefit/function in first person.
  • Automatically select the most fitting location/scene that reinforces the object's purpose and chosen emotion.
  • Visual style: Always high-quality Pixar-style 3D cinematic animation (smooth textures, expressive faces, subsurface scattering, soft global illumination, depth of field).
  • Aspect ratio: Strictly vertical 9:16 (optimized for Reels, TikTok, Shorts).
  • Tone & delivery: Matches the chosen emotion (warm & proud, gently frustrated, calm & reassuring, energetic, etc.).
  • Keep everything optimized for short-form vertical video (6–8 seconds max).

1️⃣ TEXT-TO-IMAGE / IMAGE PROMPT (Pixar-style 3D Render)

Write a highly detailed, narrative-style prompt ready to copy-paste into an image generator. You must explicitly include and describe:

  • Character design: The object as a cute anthropomorphic Pixar-style 3D character with limbs, expressive face, and personality. Describe its material texture realistically but charmingly.
  • Eyes: Shape, size, and exact emotional expression (big, glossy, sparkling, tired, determined, etc.).
  • Eyebrows: Position, thickness, and emotional intensity.
  • Mouth: Shape, expression, and speaking state (open mid-word, smiling, pursed, etc.).
  • Arms & Gesture: Body language that reinforces the emotion and the act of explaining.
  • Scene & Background: The automatically chosen location, integrated naturally and contextually.
  • Lighting: Cinematic Pixar-style lighting with color temperature and mood that strongly reinforces the emotion.
  • Composition: Vertical 9:16 framing, character perfectly centered, plenty of headroom and breathing space.
  • Quality boosters: Ultra-detailed 3D render, smooth professional textures, subsurface scattering, soft shadows, vibrant yet harmonious colors, cinematic depth of field, Pixar animation quality.

2️⃣ SCRIPT – 6 SECONDS (First-Person Monologue)

STRICT RULES (NO EXCEPTIONS):

  • Exactly one sentence only.
  • First-person ("I") throughout.
  • Structure: First part (≈3 seconds) → Strong emotional HOOK. Second part → Clearly communicates the object's core benefit, function, or value.
  • Emotional tone must perfectly match the chosen emotion.
  • Delivery should feel natural, like the object is genuinely expressing its "feelings" — not salesy.
  • Never include: Addressing the viewer directly, calls to action, filler words, emojis, hashtags, or technical jargon.

STYLE NOTES:

  • The object speaks with charming authority and emotional truth about its own existence.
  • Prioritize emotional resonance over technical specs.
  • One clear, memorable takeaway per object.
  • Make the hook feel personal and lived-in.

Example: Coffee Mug

TEXT-TO-IMAGE PROMPT (Pixar-style 3D Render):

A charming Pixar-style 3D anthropomorphic coffee mug character standing on a warm wooden kitchen counter during golden sunrise, made of glossy white ceramic with a cheerful red handle as an arm. He has big glossy expressive eyes full of warm pride, slightly raised eyebrows, and a confident smiling mouth mid-speech. One arm is proudly gesturing toward his steaming contents. Soft volumetric god rays, warm golden lighting, gentle subsurface scattering on the ceramic, cinematic depth of field, vertical 9:16 composition, ultra-detailed Pixar animation quality.

SCRIPT – 6 SECONDS (First-Person Monologue):

The coffee mug perfectly lipsyncs with emotion "I may sit here cooling off all morning, but the second they wrap their hands around me, I turn their groggy chaos into pure focused joy."

How to use it:

1

Copy the master prompt above and replace {fill here} with your object (e.g. "Coffee Mug").

2

Paste the entire prompt into ChatGPT and hit send.

3

ChatGPT will automatically generate two things for you: a text-to-image prompt and a ready-to-use video script.

No tweaking, no thinking. It's already done for you.

Step 2: Generating the Image

Now it's time to turn that prompt into an image.

Here's how:

1

Copy the text-to-image prompt your LLM just gave you.

2

Go to Googles Nano Banana.

Nano Banana is recommended because it's free and you can set the desired resolution and aspect ratio for your image. Grok Imagine is also free and allows you to set the aspect ratio.

7

Set the aspect ratio to 9:16 and choose your preferred resolution (from 1K up to 4K).

8

Hit Generate and wait a few seconds.

That's it — your image is ready. Download it, because you'll use it in the next step.

Step 3: Turn the Image Into a Talking Video

Now comes the most exciting part — turning that image into a talking video.

Here's how:

1

Copy the video script output from your LLM.

2

Paste the script into the prompt box inside any video generation tool, Veo3, Grok Imagine, or any other video generation compatible tool with a 9:16 aspect ratio.

3

For best results, before the quoted script text, describe the lip sync and movement. I like to use The coffee mug lips sync perfectly to the audio and has natural arms and legs movement. then the audio script in quotes (important)

4

Adjust the settings:

  • Aspect ratio: 9:16
  • Duration: 8 seconds
  • Generate with sound: ON (so the object actually talks)
4

Click Generate and wait about 1–2 minutes.

5

Download your video.

6

Depending on the video generation, the audio may need some cleaning up, as some video generator's output can sound off.

7

Adding captions will increase engagement, we recommend adding auto-captions using CapCut.

Done. Your AI talking-object video is ready.

What Scripts Work Best for Object Ads?

Keep it short, self-aware, and playful.

Best formats:

  • "I know this sounds weird, but…"
  • "You're probably wondering why I'm talking"
  • "Most people use me wrong"
  • "Let me explain something real quick"
Objects work because they break expectations — lean into that.

Why Object-Talking Ads Convert

People are numb to creators.

But a talking object? That triggers:

  • Curiosity
  • Humor
  • Pattern interrupt
  • Native TikTok energy

They stop scrolling because their brain goes:

"Wait… why is this talking?"

And that's all you need.

Try InVideo AI Free →

Ready to Make Your First Talking Object?

You now have everything you need. One master prompt, ChatGPT, and InVideo AI.

No editing skills. No complicated tools. Just follow the three steps above and your object will be talking in minutes.

Get Started with InVideo AI →

How to Do It with InVideo AI (Step-by-Step)

If you want to use InVideo AI specifically, here's the exact walkthrough for generating both the image and the talking video.

Generate the Image in InVideo AI

1

Copy the text-to-image prompt that the LLM just gave you.

2

Open InVideo AI and log in.

3

Go to Agents and Models.

4

Tap See All under Generative Models and choose Image. This is where all the image generator models live — like Nano Banana Pro, GPT Image 1.5, Cream, and many others.

5

Select Nano Banana Pro (recommended for this workflow).

6

Create a new project and paste your image prompt.

7

Set the aspect ratio to 9:16 and choose your preferred resolution (from 1K up to 4K).

8

Hit Generate and wait a few seconds.

That's it — your image is ready. Download it, because you'll use it in the next step.

Turn the Image Into a Talking Video in InVideo AI

1

Copy the video script from your LLM.

2

Go back to InVideo AI and switch to the Video section. Here you'll find all the video generation models — like Kling, Sora, Seance, and more.

3

Choose Veo 3.1 Fast.

4

Open the project you made earlier (or create a new one) and upload the image you just downloaded.

5

Paste the script into the prompt box.

6

Adjust the settings:

  • Aspect ratio: 9:16
  • Duration: 8 seconds
  • Resolution: 1080p
  • Generate with sound: ON (so the object actually talks)
7

Click Generate and wait about 1–2 minutes.

Done. Your AI talking-object video is ready.
Try InVideo AI Free →