Launch Directory/AI Agents/Abdul Șhakoor

Abdul Șhakoor · Launch Video Breakdown: Hook, Pacing & Motion Design

Decart just dropped Lucy 2.5 and it completely changes what live video can do. This is their Live AI model that edits video as it streams, no post production needed. > add, replace, or remove anyth…

AI AgentsLaunchJuly 16, 2026@abxxai

Website View on X

9.2 /10

Hook score

Immediate Product Demonstration

Cuts / min

0:22

Runtime

209K

Views

Likes

Reposts

Replies

Reusable motion template

Live AI Real-time Effect Showcase

Live AI Technology Demonstration & Comparison

Adapt it.

Replace the main video source with a new video of a presenter speaking.
Implement a custom audio processing module (e.g., Web Audio API with AI model integration) to apply live voice modulation effects.
Ensure the picture-in-picture component displays the original, unmodulated audio/video feed.
Update the text overlay to 'VocalForge' and adjust its position if necessary.
Synchronize audio effects with the presenter's speech, demonstrating different modulations sequentially.

Source: X / Twitter launch · View original post

Launch overview & key findings

A 22-second live ai technology demonstration & comparison launch video by Abdul Șhakoor, unveiled on July 16, 2026 with over 209K views on X. Evaluated at a 9.2/10 hook score with an editing tempo of 3 cuts per minute, the production demonstrates Abdul Șhakoor's Decart just launched Lucy 2.5, a groundbreaking Live AI model that edits video as it streams, eliminating the need for post-production. The video showcases a woman's face being dynamically distorted and manipulated in real-time, with a smaller picture-in-picture providing a clear comparison to the original feed. The motion style is a raw, continuous live demonstration, emphasizing the AI's seamless and interactive capabilities without any cuts or elaborate effects.

0:00 · The Hook · AI-Powered Face Manipulation Teaser

0:00 / 0:00

The Hook
AI-Powered Face Manipulation Teaser
“(No spoken dialogue — silent video with ambient office sounds implied by visual context.)”
On screen
+ decart
Camera
Static wide shot of a woman at a desk, with a smaller picture-in-picture showing the original, unedited feed.
Motion
Real-time AI facial distortion applied to the main video feed, contrasting with a clean reference feed.
Product Reveal
Dynamic Facial Feature Exaggeration
“(No spoken dialogue — silent video with ambient office sounds implied by visual context.)”
On screen
+ decart
Camera
Main camera remains static, focusing on the woman's face as AI effects dynamically alter her features.
Motion
Live AI model applying exaggerated facial deformations (e.g., puffed cheeks, stretched mouth) in real-time.
Feature Teaser
Interactive AI Facial Control
“(No spoken dialogue — silent video with ambient office sounds implied by visual context.)”
On screen
+ decart
Camera
Static wide shot, with a male colleague appearing in the background, observing the AI effects.
Motion
The AI model responds to the woman's gestures (e.g., touching her face), creating interactive and fluid distortions.
Call to Action
Seamless Live AI Editing Demonstration
“(No spoken dialogue — silent video with ambient office sounds implied by visual context.)”
On screen
+ decart
Camera
The main camera remains static, capturing the woman's continued interaction with the live AI effects.
Motion
Continuous, smooth application of AI-driven facial manipulation, showcasing the model's real-time capabilities without lag.

Synthesizable in Remotion · key components of the live-ai-demo-showcase template.

LiveFaceDistortion

A subject's face is dynamically distorted and exaggerated in real-time, responding to subtle movements or gestures. Features like cheeks or lips inflate and deflate fluidly.

+ Code construction recipe

Utilize Remotion's `<Video>` component with a custom WebGL shader or Canvas filter. Implement a face-tracking library (e.g., MediaPipe) to get facial landmarks, then use `interpolate()` to map landmark changes to shader uniform values for distortion effects (e.g., 'bulge' or 'pinch' filters).

PictureInPictureReference

A smaller, unedited video feed is displayed concurrently with the main, processed video, providing a clear comparison of the applied effects.

+ Code construction recipe

Render two `<Video>` components. Position the main video full-screen and the reference video in a smaller, rounded-rectangle container using CSS `position: absolute`, `border-radius`, and `box-shadow`. Ensure both videos are synchronized using `useCurrentFrame()` and `continueRender()`.

RealtimeTextOverlay

A subtle, static text overlay (e.g., company name) is present in the corner of the screen throughout the video.

+ Code construction recipe

Use a `<AbsoluteFill>` with a `<Text>` component. Apply `position: absolute`, `top`, `right` CSS properties for placement. Use a simple, clean font and a semi-transparent background if needed for readability.