AI cartoon with 20 million views: 14 steps and 3 critical nuances

We will walk through the entire process from the initial concept to final export: storyboarding, character generation, image creation, animating frames, sound design, editing, and ad integration. This approach works for both personal blogs and businesses looking to regularly release animated videos without a film crew.

Why some videos go viral while others do not

A viral video is almost always built on one formula: a familiar beginning, an unexpected twist, an emotional response, and a finale that reinforces that emotion.

For example, in a video that gathered around 20 million views, a mother is feeding her child. Everything is recognizable and calm. Then, instead of eating, the child starts playing with the pasta—making a hairstyle and a hat out of it. Expectations are broken, a smile appears, and the viewer watches until the end.

Important: a hook in the first second is not mandatory. Sometimes it is better to start with a familiar everyday scene and introduce the surprise slightly later. People appreciate content that evokes emotion: surprise, tenderness, laughter, or relatability.

This is exactly why AI cartoons featuring animals, everyday situations, and cute characters perform so well. An AI cat cartoon is nearly a ready-made format: cats are understandable, charming, easily placed in funny situations, and work excellently in advertising.

We recorded a comprehensive free tutorial with Boris Zarubin, where he demonstrates step-by-step how to create animated videos that quickly go viral and stick in memory. He also shared advice on monetization and key nuances to ensure virality.

Step 1. Come up with an idea, not just a pretty picture

If you are looking for how to create a cartoon using neural networks, start not with the AI, but with the idea. Beautiful generation without meaning is just a set of frames. A strong idea turns a clip into a story.

Any idea has three parts:

Even a 15–30 second clip needs a mini-story. Instead of just “a cat pouring coffee,” try “a cat making coffee, licking some milk, and still serving the drink beautifully.” That’s the kind of cartoon cat video people want to watch until the end.

Quick exercise: take a common situation from your niche and add one unexpected action by the character. For example: “barista makes coffee” becomes “cat barista makes coffee and tastes the milk straight from the pitcher.” This instantly adds personality and an emotional hook.

Step 2. Choose a format for social media

If you want to generate an AI cartoon for Reels, Shorts, or TikTok, set the technical parameters right away. For a short promotional video, consider:

• Vertical 9:16 format.

• Duration of about 30 seconds.

• 4–6 scenes.

• A clear story without unnecessary dialogue.

• Scene sounds instead of long voiceovers.

In the barista cat example, we chose 30 seconds, a vertical format, and five scenes. This is enough to introduce the character, show the action, highlight the unique feature, and end with a brand frame.

Don’t confuse duration with the number of scenes: one setting error can cost you extra tokens and time. Always check parameters before starting generation.

Step 3. Prepare branding in advance

If you’re creating a video for business, upload your logo or brand guide before generation. It’s better to use PNG files with transparent backgrounds—this makes it easier to integrate the brand into scenes.

What to prepare in advance:

1. Logo.

2. Brand colors.

3. Product: coffee, packaging, app, service, or item.

4. Desired brand placement in the frame.

5. Video style.

The 3D Pixar style works great for a modern coffee shop. However, a “vintage fairy tale” style can subtly alter the equipment, interior, and atmosphere: the AI will adjust the entire scene to match the chosen aesthetic. Therefore, the style should support the story, not disrupt it.

Step 4. Work with a storyboard, not chaotic prompts

A storyboard is your visual script: scene by scene, with clear actions and objects in the frame. An AI storyboard saves a huge amount of time: you don’t need to manually transfer prompts between services, copy descriptions, or regenerate frames dozens of times.

In Lavanda GPT —a service that acts as an AI aggregator—there is a dedicated Storyboard tool. It helps you go from idea to finished video in one window: write the concept, get a script, generate frames, animate them, and export the video.

Why this is convenient:

• The idea immediately turns into a structured script.

• Prompts for frames are created automatically.

• Characters can be saved and reused.

• Branding is added to all relevant scenes.

• No need to switch between dozens of services.

If you are looking for a systematic way to create animations with AI, start with a storyboard. It is your production pipeline.

Step 5. Write a prompt as a brief for a film crew

Imagine the AI as your film crew: screenwriter, animator, artist, and cinematographer. You need to explain the task as clearly as you would to a human contractor.

Don’t write: “make a cartoon about a cat.” Be specific. Here is an example of a working prompt we generated for a video featuring a barista cat:

Create a 30-second vertical animated ad consisting of 5 scenes.

Character: a white, fluffy cat barista who is kind and slightly humorous.

Location: a cozy, modern coffee shop.

Style: 3D Pixar animation.

Branding: the coffee shop logo appears on the sign, cups, and packaging.

Plot:

1. The cat approaches the espresso machine.

2. Coffee grounds fill the portafilter.

3. The cat steams milk, tastes it, and continues working.

4. The cat pours milk into the cup with artistic flair.

5. The cat serves the finished coffee and looks at the camera.

Important: maintain character consistency, logical object placement, and natural physics. Do not add unnecessary dialogue. Exclude music; keep only ambient scene sounds.

Upload the prompt here: 

The more precise the description, the fewer AI-generated errors you will encounter. If the cat is holding a pitcher, specify how: “Holding the pitcher with front paws, only paws visible.” If the espresso machine is important, explain where the steam wand is located and where the milk is being poured.

Step 6. Add helper prompts to reduce errors

A helper prompt is an additional instruction for the AI scriptwriter. It does not eliminate errors entirely but significantly reduces their frequency. Write it in the same window as your main script, simply adding one or two paragraphs at the end.

You can include the following in the helper prompt:

• Maintain scene logic.

• Do not detach objects from their context.

• Show natural character movements.

• Pay attention to the placement of equipment and objects.

• Avoid adding unnecessary details that disrupt the frame.

This is especially important if you are creating AI-generated animations for advertising: an error with a cup, coffee machine, or product immediately undermines the sense of quality. And every extra regeneration costs time and money.

Step 7. Review the storyboard before generating video

When the script is ready, don’t rush to click “Create video.” Read through the storyboard first.

Check:

• Whether the story is clear from the first to the last frame.

• Whether the sequence of actions is correct.

• Whether the product and brand are mentioned in the right scenes.

• Whether voiceover is needed or if sound effects are sufficient.

• Whether there is a “hook” that makes viewers want to watch the clip again.

If a scene doesn’t match your vision, edit it immediately by clicking the pencil icon. This way, you fix the problem at the text stage rather than having to redo the generated video later.

Step 8. First, create a consistent character

Before generating scenes, create the main character. This is a key step for anyone who wants to understand how to make an AI animation rather than a random set of clips. Start by generating a cat, person, or brand mascot. Check their appearance, color, proportions, and personality. Then save the character and use them in all scenes.

The character is generated automatically based on the description of their appearance in the script. You just click the “Magic Wand” here:

Why this is important:

• The character remains recognizable.

• You can produce a series of videos.

• The character becomes the brand’s mascot.

• There is no need to re-describe their appearance every time.

A barista cat is a great example: the same character can brew coffee, take orders, get into funny situations, and promote various café products.

Step 9. Choose models based on the task

Generating photos and images is only the first stage. Next, you need to bring the frames to life. And here, it is not necessary to use the most expensive model for every scene. You can select the appropriate one directly below the character image. 

For simple scenes, choose a more affordable model. If physics, objects, fluid motion, or complex logic are important in the frame, use a more powerful model. This way, you control your budget and get predictable results.

Step 10. Start with low resolution

Use a lower resolution for drafts. In our example, we used 480p, which is enough to check composition, characters, objects, and scene logic. Once the storyboard is approved, increase the quality. For social media, 720p is usually more than sufficient.

This simple rule saves tokens: first find a successful idea and composition, then pay for final quality.

Step 11. Look for AI artifacts in every frame

Before animating images, carefully examine each starting frame. Small errors reveal AI-generated videos and undermine viewer trust.

What to check:

• Physics: Is the cup floating? Is the coffee machine positioned correctly?

• Objects: Are there extra pens, wires, fingers, or paws?

• Character: Have the color, face, clothing, or proportions changed?

• Logic: Where is the milk pouring from? Where is the character looking? What are they holding?

• Brand: Has the logo or product disappeared?

For example, in our third scene, the milk is being frothed with some unclear stick, and the steam wand is turned away (in the upper right corner). Also, the cat’s paws became striped in this same scene. 

Such details are easy to fix by tweaking the prompt directly below this frame: specify where the pitcher is, how the wand is positioned, what exactly the cat is holding, and what color the paws should be.

Do not try to rewrite the entire prompt and script because of one small detail. Edit the specific scene and regenerate only that part.

Step 12. Bring shots to life and trim the excess

After reviewing the images, start generating video by scene. Then assemble the clip and watch it as an ordinary viewer would.

Here is what you can almost always cut using the built-in video editor for each scene:

• Long static moments.

• Repetitions of the same action.

• Extra seconds after an action has finished.

• Shots where everything is already clear.

• Unsuccessful fragments that are easier to regenerate.

In the cat example, a drawn-out shot was trimmed to about two seconds, while the final shot of the cat was left intact: the character looks at the camera and serves coffee, which enhances the emotional impact.

Step 13. Make sound part of the story

If you are combining several fragments into one video, it is better to generate scenes without music. Otherwise, each segment will have its own track, creating awkward transitions and requiring you to redo the final audio.

The best order is as follows:

1. Keep the scene’s native sounds: the coffee machine, steam, pouring, footsteps, and the character’s reactions.

2. Assemble the video.

3. Add one consistent music track during editing.

4. Check volume levels between fragments.

5. Enable loudness normalization when exporting.

In Lavanda GPT Sound design can already be integrated with the video: coffee machine noise, steam, pouring—everything is created at once. It’s easier to add a single music track for the entire clip in the editor.

Step 14. Export and watch the video from the viewer’s perspective

Before publishing, turn on the sound, watch the entire video, and honestly ask yourself: “Would I share this with a friend?”

Final checklist:

• Is the idea clear within the first few seconds?

• Is there emotion and an unexpected detail?

• Does the logic hold up in the middle?

• Is the brand visible without disrupting the story?

• Are there any obvious physics or object errors?

• Is it easy to watch without sound?

• Do you want to finish watching and share the video?

Don’t become an endless perfectionist. For a cute animated ad, viewers will forgive minor cartoonish conventions. But if you’re creating a serious cinematic video, the standards for details should be higher. 

So, let’s take another look at the finished video, now understanding how we achieved each transition, sound, and movement: 

How to integrate advertising so the video doesn’t “die”

Advertising should be part of the story, not just a five-second insert. If a cat barista is making coffee, it makes sense to show the café, the drink, the cup, the equipment, or the brand.

A logo on the mug looks natural

Rules for native integration:

• The product supports the plot rather than interrupting it.

• A logo lives in its environment: signage, cups, packaging, and phone screens.

• Avoid mixing incompatible elements: sudden investment ads in a cozy family scene can alienate your audience.

• Study your niche: see which videos are working for competitors and adapt successful mechanics to your brand.

For example, a barista cat video can simultaneously entertain, gather views, and sell coffee.

How to turn one video into a revenue system

If you are looking to earn from videos and views, do not rely on a single hit. One video is a test. A system requires consistency.

Here is a working approach:

1. Release 10–20 short AI animations with different ideas.

2. Check which videos drive watch time, shares, comments, and clicks.

3. Identify 2–3 formats that perform better than the rest.

4. Create a series featuring one recognizable character.

5. Integrate your product into the most successful scenarios.

6. Build a portfolio and offer videos to businesses.

Do not wait passively for orders. Create a portfolio, post your work on freelance platforms, join relevant communities, and pitch your services directly to businesses. Packages work especially well: for example, 10–15 videos or a full series for a brand.

Also, avoid using famous cartoon characters without significant transformation. Create your own unique character so you can develop it, produce a series of videos, launch merchandise, and build a recognizable brand.

You might also like:

Ready-made template for a viral AI animation 

Copy this template, customize the details for your product, and you can start creating AI animations instead of just dreaming about them.

Create a 30-second vertical animated ad consisting of 5 scenes.

Character: [describe the character: appearance, color, personality, clothing, details].

Location: [setting and atmosphere].

Style: [e.g., 3D Pixar].

Brand: [name, logo, product, and where it appears].

Plot:

1. First scene: [a familiar everyday situation].

2. Second scene: [the character starts interacting with the product].

3. Third scene: [an unexpected or funny detail].

4. Fourth scene: [the product enhances the emotion or solves the problem].

5. Fifth scene: [final shot with the brand and a clear conclusion].

Important: maintain character consistency, logical object placement, natural physics, and movement. Avoid unnecessary dialogue if the story is clear visually. Do not add music; use only natural scene sounds.

Now you have more than just theory—you have a working process: idea → storyboard → character → image generation → review → animation → sound → editing → ad. Follow this workflow, create many short videos, and explore AI video trends. Eventually, one of them could become the viral hit that brings you both audience and customers.

After registration, you will receive free tokens for testing