Guide: how to create AI videos that get millions of views

3 Proven Combos | Ready-to-Use Prompts | Step-by-Step Instructions from Idea to Upload

If short-form video is used as the top of a funnel that moves viewers into Telegram, it helps to separate publishing from the next stage of audience management. Telegram Expert can serve as the Telegram software layer for multiple accounts, channels and chats, different types of broadcasts, and auto-replies. This keeps content production separate from the Telegram-side workflow.

Concept

Take a familiar household object and use an AI tool to add realistic eyes and a mouth. The object shares its pain: complaining about wear and tear, accumulated bacteria, or its hidden dangerous nature. It holds viewer attention through an eerie uncanny valley effect and pattern-break storytelling.

ChatGPT Prompt

You are an AI creator of viral visual content specializing in realistic anthropomorphic objects.

Your task is to take the object [OBJECT] and create TWO SEPARATE PROMPTS in a single cohesive style: unsettling realism, hidden truth, and viral AI animation.

1. Image Generation Prompt

Create a prompt to generate a photorealistic image of the object.

Requirements:

• The object must be anthropomorphic: small realistic eyes, a mouth, minimal arms or a subtle hint of a gesture

• The object itself remains the primary subject of the frame

• The environment must be realistic and typical for this object

• Must show clear signs of use: scuffs, dirt, wear, micro-cracks, and realistic textures

• The object's emotional state should be visible: exhaustion, anxiety, sadness, honesty, fatigue

• Visual style: photorealistic with barely noticeable cartoonish facial traits

• Scene mood: slightly unsettling, honest, cautionary, revealing a hidden truth

• Lighting, textures, and details must look as natural as possible

• The image must be cinematic and viral in presentation, without being overly cartoonish

• Prompt language: English

2. Animation Prompt

Create a separate prompt to animate the generated image.

Requirements:

• Animation should be minimal and realistic

• Movements: rare blinking, subtle facial expressions, faint gestures, natural swaying

• The object speaks in English, with proper lip synchronization

• Speech script (220–260 characters) must explain a hidden problem with the object (why it is harmful, needs replacement, or requires cleaning)

• Tone: informative, calm, tired, free of humor or shouting

1. Script: Send the prompt above into ChatGPT. When prompted, specify any item (e.g., "Dishwashing sponge"). Receive two English prompts (one for the image, one for animation).

2. Image: Head to an image generator (such as Midjourney or Google Flow), paste the first prompt, and save the generated image.

3. Animation: Open an AI video animation tool (Grok, Luma, Kling, or Runway). Upload the image, paste the second prompt, and generate a video where the object breathes and moves its lips.

4. Voiceover: Take the speech text (generated in GPT's second prompt) and run it through ElevenLabs to create a realistic voiceover.

5. Editing: Combine the video and audio track in CapCut, generate auto-captions, and the video is ready for publishing.

Concept

Unearth wild historical facts and present them as a high-stakes thriller. To keep production costs low and stand out visually, generate scenes in a 16-bit retro gaming style featuring anthropomorphic cats instead of humans.

Stage 1: Idea Generation Prompt

ROLE: You are a creative producer for a viral TikTok/YouTube Shorts channel. You specialize in finding real historical events that sound so wild viewers think, "Did that actually happen?!" Your style is shock, absurdity, incredible coincidences, and crazy exploits.

TASK: Come up with 10 video ideas (under 60 seconds). Each topic must be based on real facts and delivered with a clickbait hook.

Requirements:

• Instantly grab attention and provoke surprise

• Contain a strong emotional trigger

• Avoid sounding like a textbook

• Offer variety (scams, strange deaths, absurd laws, etc.)

OUTPUT FORMAT: For each idea, provide a Title and a Pitch (one sentence summarizing the main conflict).

Stage 2: High-Impact Scriptwriter Prompt

ROLE: You are a top-tier scriptwriter for TikTok, turning real events into short, cinematic stories.

TASK: Write a script for a short animated video on the topic: [INSERT TOPIC FROM PREVIOUS STEP].

PARAMETERS:

• Length: 1000–1100 characters (strictly timed for a 60-second voiceover).

• Format: Break into short paragraphs.

STRUCTURE:

1. Hook (0–5 sec): Shocking opening sentence.

2. Context/Pain Point (5–15 sec).

3. Twist/Action (15–40 sec): A clever move, risk, or scheme.

4. Viral Fact (40–50 sec).

5. Ending (50–60 sec): High-impact conclusion.

STYLE: Energetic, sharp, short sentences, free of dry corporate language.

Stage 3: Scene Generation Prompt (Pixel Art)

ROLE: You are an AI visualizer. I will give you sentences from the script, and your task is to translate them into precise scene descriptions in English.

FIXED STYLE PARAMETERS:

• Style: Retro 16-bit pixel art graphics, SNES RPG cutscene aesthetic, visible pixels, limited color palette.

• Characters: Anthropomorphic cats in period-accurate historical clothing.

• Atmosphere: Serious, dramatic. No text on image.

OUTPUT FORMAT: [Character description] [Main action] [Clothing / weapons / details] in [setting]. No text. Style: retro 16-bit pixel art animation, SNES RPG cutscene aesthetic.

Break my script into 5-6 key scenes and generate prompts for each. Script: [INSERT SCRIPT FROM PREVIOUS STEP]

1. Script Generation: Run Prompts 1, 2, and 3 sequentially through ChatGPT to get a completed script and a set of visual prompts.

2. Art: Generate static images using your prompts in an AI generator capable of rendering clean pixel art.

3. Animation: Upload static images to a video AI generator with the instruction: "Animate this scene in the same 16-bit pixel art style with smooth movements."

4. Final Assembly: Generate voiceover via ElevenLabs and assemble the clips in your video editor.

Concept

The "what happens to the body if..." niche consistently draws massive engagement. We visualize physiological processes using hyper-realistic 3D models of translucent skeletons. It looks high-tech, frightening, and delivers exceptionally high audience retention.

Stage 1: Topic Generator & Script

• Step 1 (Ideas): Give me 5 clickbait video ideas exploring the limits of the human body in formats like: "What happens if...", "Could you survive...".

• Step 2 (Script): Write a 60-second script (1000 characters), divided into exactly 14 stages (seconds, minutes, days). Style: A dark chronicle of physical breakdown. Each stage is more intense than the last. Topic: [YOUR TOPIC]

Stage 2: 3D Visualization Prompt (Images)

You are a 3D visualization engineer. Break my script into scenes and create English prompts for image generation.

Unified Style:

• Hyper-realistic medical 3D rendering, Unreal Engine, Octane Render

• Transparent/translucent human skeleton made of glossy plastic

• Internal organs visible inside, large realistic moist eyes

• Cinematic lighting, vertical 9:16 aspect ratio.

Script: [YOUR SCRIPT]

Stage 3: Smooth Animation Prompt

Analyze the script and generated image prompts. Write instructions (prompts) for an image-to-video AI to animate these pictures.

Movement rules: Slow, organic motion, heavy realistic breathing, faint eye movements. Static camera. Output instructions in English.

For auto-uploading, managing large account networks, and safely bypassing spam filters, professional teams rely on the Aether Tools ecosystem. Depending on your target platforms, you can leverage specialized software:

• Aether YT - for YouTube channel automation, scheduling, and metadata randomization (tags, titles, descriptions).

• Aether IG - for Instagram network management (Reels posting, account warm-ups, and browser-emulated account actions).

Using the Aether Tools ecosystem allows you to configure unique digital fingerprints, proxy connections, and action timing, transforming AI video generation and posting into a fully automated, scalable, and protected business process.