0:00 / 0:00

Claye Ai

@claye.ai

INSTAGRAM · 2026-03-21Source

Remix This

Recreate with Kling 3

Make your own AI viral video

Prompt

GLOBAL LOCK: 
Subject: A young woman of South Asian descent, olive skin tone, long wavy dark brown hair with subtle highlights, expressive brown eyes, natural makeup. 
Wardrobe: Left side wears a white ribbed knit crewneck sweater; Right side wears a black ribbed knit crewneck sweater. 
Environment: A professional podcast studio setting. A polished dark wooden table in the foreground. A professional condenser microphone with a red shock mount is mounted on a small black tripod in the center. 
Lighting: Split-screen lighting logic. Left side has a deep purple and magenta ambient glow in the background. Right side has a deep blue and cyan ambient glow. Soft key lighting on the subject's face from the front. 
Camera: Static Medium Close-Up (MCU), eye-level angle, shallow depth of field with a blurred studio background. 
Speech Style: Professional, rhythmic, direct-to-camera delivery. Clear articulation with a warm, helpful tone.

[00:00–00:02]
Subject: The woman is centered in the split screen, looking directly at the camera with a friendly smile.
Action: She begins speaking the intro. Subtle head tilt.
Framing: MCU, split screen.
Text Overlay: "Free VS Paid ai tools" appears at the top center. "Paid" on the left, "Free" on the right.
Speech: "Free versus paid AI tools. Let's compare."
Lip-sync: High strictness.

[00:02–00:05]
Subject: Subject continues speaking, maintaining eye contact.
Action: Natural blinking and slight hand gestures near the microphone.
Environment: "image generation" text appears at the top.
Logos: Midjourney logo appears on the left ("Paid"), Google Gemini logo on the right ("Free").
Speech: "Midjourney, Google Gemini."
Lip-sync: High strictness.

[00:05–00:08]
Subject: Subject maintains the same pose.
Environment: "AI chat assistants" text appears.
Logos: ChatGPT logo on the left, DeepSeek logo on the right.
Speech: "ChatGPT, DeepSeek."
Lip-sync: High strictness.

[00:08–00:11]
Subject: Subject maintains the same pose.
Environment: "AI video editing" text appears.
Logos: Kling AI logo on the left, Artflow logo on the right.
Speech: "Kling 1.0, Artflow."
Lip-sync: High strictness.

[00:11–00:14]
Subject: Subject maintains the same pose.
Environment: "AI ads" text appears.
Logos: Canva logo on the left, Gemini Pomelli logo on the right.
Speech: "Canva, Gemini Pomelli."
Lip-sync: High strictness.

[00:14–00:17]
Subject: Subject maintains the same pose.
Environment: "AI video generation" text appears.
Logos: Google Veo 3.1 logo on the left, Meta AI logo on the right.
Speech: "Veo 3.1, Meta AI."
Lip-sync: High strictness.

[00:17–00:20]
Subject: Subject maintains the same pose.
Environment: "voice cloning" text appears.
Logos: ElevenLabs logo on the left, Fish Audio logo on the right.
Speech: "ElevenLabs, Fish Audio."
Lip-sync: High strictness.

[00:20–00:22]
Subject: Subject maintains the same pose.
Environment: "AI avatars" and "Lipsync videos" categories flash quickly.
Logos: HeyGen, Wondershare Virbo, Infinite Talk, and Wand 2.2 logos appear and swap.
Speech: "HeyGen, Wondershare Virbo, Infinite Talk, Wand 2.2."
Lip-sync: High strictness.

[00:22–00:25]
Subject: The background blurs further and darkens.
Action: The subject is still visible but the focus shifts to a central graphic.
Visual: A cluster of the previously mentioned AI logos floats around a central purple geometric icon.
Text Overlay: "comment AI" with a heart icon and a small profile picture of the creator.
Speech: "Want all these tools with links? Comment AI and I'll send the full list."
Lip-sync: High strictness.

NEGATIVE PROMPT: 
Visual: Unnatural facial warping, flickering background lights, inconsistent hair movement between cuts, extra fingers, distorted microphone shape, blurry logos, text spelling errors, low-resolution textures, robotic or stiff body movement.
Speech: Robotic monotone, misaligned lip-sync, background hiss, popping 'p' sounds, unnatural pauses mid-word, metallic voice artifacts, inconsistent volume levels.

SPEECH PACK:
[00:00-00:02] "Free versus paid AI tools. Let's compare."
TAKE_A: (Energetic) Free versus paid AI tools! Let's compare.
TAKE_B: (Professional) Free versus paid AI tools. Let's compare.
TAKE_C: (Casual) Free vs paid AI tools... let's compare.

[00:02-00:20] "Midjourney, Google Gemini. ChatGPT, DeepSeek. Kling 1.0, Artflow. Canva, Gemini Pomelli. Veo 3.1, Meta AI. ElevenLabs, Fish Audio. HeyGen, Wondershare Virbo, Infinite Talk, Wand 2.2."
TAKE_A: (Rapid fire, rhythmic) [Tool Name], [Tool Name]. (Pause) [Tool Name], [Tool Name].
TAKE_B: (Steady pace) [Tool Name] versus [Tool Name].

[00:20-00:25] "Want all these tools with links? Comment AI and I'll send the full list."
TAKE_A: (Direct) Want the full list? Comment AI and I'll send it over!
TAKE_B: (Helpful) Comment AI below and I'll send you all the links.

How claye-ai Made This Free vs Paid AI Tools Comparison AI Video — and How to Recreate It

This case study analyzes a high-performing "Free vs. Paid" AI tool comparison video by @claye.ai. The video utilizes a cinematic split-screen aesthetic featuring a consistent AI-generated female avatar. By positioning "Paid" tools on the left (warm purple/pink lighting) and "Free" alternatives on the right (cool blue lighting), the creator establishes a clear, visually satisfying framework for delivering high-density information. This "curated list" format is a proven save-magnet for indie creators looking to build authority in the AI and productivity niches.

What You’re Seeing: A Visual Breakdown

The video features a young woman with olive skin and long wavy brown hair, acting as a professional host. She is seated at a wooden desk with a high-end condenser microphone positioned centrally. The wardrobe changes slightly between the split screens—a white sweater on the left and a black sweater on the right—reinforcing the "Paid vs. Free" contrast.

Shot-by-Shot Breakdown

Time Range	Visual Content	Shot Language	Lighting & Tone	Viewer Intent
00:00–00:02	Avatar introduces the "Free vs Paid" concept.	Medium Close-Up (MCU), Static.	Dual-tone: Purple (L) / Blue (R).	Hook: Immediate value proposition.
00:02–00:18	Rapid-fire tool comparisons (Midjourney vs Gemini, etc.).	Split-screen MCU, lip-synced.	Cinematic, shallow depth of field.	Retention: High-density info delivery.
00:18–00:22	Avatar gestures toward floating logos.	MCU with motion graphics.	Consistent studio lighting.	Reinforce Authority: Showing variety.
00:22–00:25	Call to Action (CTA) overlay with comment prompt.	Darkened background, UI overlay.	High contrast.	Conversion: Triggering comments.

Why It Went Viral: The "Utility + Aesthetic" Formula

The Content Strategy

This video taps into the "Information Arbitrage" psychology. In a rapidly evolving field like AI, users are overwhelmed by choices. By curating a "Free vs. Paid" list, the creator provides immediate relief to the viewer's "Fear of Missing Out" (FOMO). The split-screen format allows for a direct comparison that is biologically easy for the brain to process, making the content feel organized and trustworthy.

The Aesthetic Hook

Unlike typical "talking head" videos, this uses a hyper-realistic AI avatar. The visual quality is "uncannily perfect," which creates a secondary layer of curiosity: "Is this a real person or AI?" This ambiguity often drives users to the comments section to investigate, further boosting the video's engagement signals.

Platform Perspective

From an algorithmic standpoint, this video is designed for Saves and Shares. Because the information is too dense to memorize in one viewing, users "Save" it for future reference. High save rates signal to Instagram and TikTok that the content is "high value," triggering a wider distribution to the Explore page. The "Comment 'AI' for the list" CTA bypasses the "link in bio" friction, turning passive viewers into active commenters, which is a massive signal for reach.

5 Testable Viral Hypotheses

The Contrast Hypothesis: Using distinct color palettes (Purple vs. Blue) for different categories increases visual retention by 30% compared to a single-tone background.
The "Save-Magnet" Hypothesis: Presenting more than 5 tools in 20 seconds forces the user to save the video to "digest" the information later.
The Identity Consistency Hypothesis: Using the same avatar across multiple videos builds "Parasocial Authority," making viewers more likely to trust the recommendations.
The Frictionless CTA Hypothesis: Asking for a specific keyword comment ("AI") instead of a general comment increases conversion rates by providing a clear, low-effort instruction.
The "Uncanny Valley" Curiosity: High-fidelity AI motion (blinking, micro-expressions) triggers a "stop-and-stare" effect as viewers try to determine the subject's authenticity.

How to Recreate: From 0 to 1

Step 1: Topic Selection & Scripting

Choose a niche where "Free vs. Paid" is a common pain point (e.g., Video Editing, Fitness Apps, Coding Tools). Script 6–8 pairs of tools.

Step 2: Character Consistency

Generate your base avatar using Midjourney. Use a prompt that specifies lighting and camera angle to ensure a professional "studio" look. Pro Tip: Use the --cref (Character Reference) tag in Midjourney to keep the face consistent across different wardrobe generations.

Step 3: Generating the Split-Screen Assets

Generate two versions of your character: one in a light outfit (for the "Paid" side) and one in a dark outfit (for the "Free" side). Ensure the background remains identical except for the color of the ambient lights.

Step 4: Animation & Lip-Sync

Use a tool like HeyGen or D-ID to animate the avatar. Upload your script or a voice recording. For the best results, use an AI voice that sounds professional and rhythmic.

Step 5: Video Generation (Motion)

If you want more natural body movement than standard lip-sync tools provide, run your Midjourney images through Kling AI or Luma Dream Machine with a prompt for "subtle head nodding and hand gestures."

Step 6: Editing in CapCut

Place the two videos side-by-side using the "Mask" or "Split" tool. Add text overlays for the tool names and logos. Ensure the cuts between tool comparisons are snappy and aligned with the audio beats.

Step 7: The "Comment Trigger" Overlay

In the final 3 seconds, add a dark overlay with a screenshot of a fake (or real) comment and a heart icon. This visually demonstrates the action you want the viewer to take.

Step 8: Publishing Strategy

Post as a Reel/TikTok. Use a "hooky" cover image that clearly shows the "Free vs Paid" text. Do not put important text in the "dead zones" (where the caption and UI buttons sit).

Growth Playbook: Distribution & Scaling

3 Ready-to-Use Opening Hooks

"Stop paying for [Tool Name]. Use these free AI alternatives instead."
"The AI gap is closing. Here is Free vs. Paid in 2024."
"I tested 50 AI tools so you don't have to. Here are the winners."

4 Caption Templates

The Value Bomb:
"Most people are overpaying for AI. 💸 I found the best free alternatives that actually work. Which one is your favorite? Comment 'LIST' and I'll DM you the links! 👇 #aitools #productivity"

Hashtag Strategy

Broad: #AI #ArtificialIntelligence #TechTrends (High reach, high competition)
Mid-tier: #AITools #ProductivityHacks #IndieCreator (Targeted interest)
Niche: #MidjourneyAlternatives #FreeAI #WorkflowAutomation (High intent)

Frequently Asked Questions

What tools make the avatar look the most realistic?

Combining Midjourney for the base image and HeyGen for the "v3.0" streaming avatar animation provides the most realistic results.

How do I keep the face the same in every shot?

Use the --cref parameter in Midjourney or a "Fixed Seed" in Stable Diffusion to maintain facial features.

Why does my AI video look "jittery"?

This usually happens due to low frame rates; try upscaling your video to 60fps using Topaz Video AI or CapCut's "Smooth Slow-mo" feature.

Is it better to use a real voice or AI voice?

For this "curated list" format, a high-quality AI voice (like ElevenLabs) often sounds more "on-brand" and professional than a bedroom recording.

How do I automate the "Comment for Link" part?

Use a tool like ManyChat to automatically trigger a DM whenever someone comments a specific keyword on your post.