
Ask a Luma Agent to extract layers from any image, then keep chatting: swap the product, restyle the background, rewrite the headline.
The Hook
“(No spoken dialogue — Sustained low-frequency drone, high-pitched synth swell, percussive glitchy riser building to a sub-bass drop/impact.)”
Product Reveal
“(No spoken dialogue — Driving electronic beat with kick drum, hi-hats, pulsating synth bassline, and a subtle high-frequency sweep.)”
Feature Teaser
“(No spoken dialogue — The driving electronic beat continues with consistent tempo.)”
Feature Teaser
“(No spoken dialogue — The driving electronic beat continues, with a significant riser effect building rapidly.)”
Call to Action
“(No spoken dialogue — Powerful sub-bass drop/impact, then the track quickly fades out with a final resonant bass tone.)”
Synthesizable in Remotion · key components of the ai-video-editor-launch template.
An initial image rapidly deconstructs into its constituent layers, which then scatter and reassemble, demonstrating the product's ability to isolate and manipulate elements.
Utilize `spring()` for individual layer movement and rotation, `interpolate()` for opacity and scale. Implement a `Sequence` to control the timing of the shatter and reassembly. SVG filters for subtle edge effects on layers.
A selected visual element within an image seamlessly transforms into a new element based on a text prompt, showcasing AI-powered content generation and editing.
Use `AbsoluteFill` for the image container. Implement a cross-fade or `scale` and `opacity` interpolation between two image components. Text input animation can be achieved with `typewriter()` effect and `spring()` for the cursor.
Multiple versions of an image, each with different text or aspect ratios, rapidly arrange themselves into a dynamic grid, emphasizing adaptability and multi-platform support.
Employ `map()` over an array of image/text data to render multiple `AbsoluteFill` components. Animate `x`, `y`, `scale`, and `opacity` properties using `interpolate()` with custom easing curves to create the grid formation and subsequent rapid cycling.
Purpose. Introduce the core capability: intelligent image deconstruction.
Execution. An image shatters into layers, demonstrating the AI's understanding of visual components.
Purpose. Showcase the ease of modifying individual layers with AI prompts.
Execution. A mountain transforms into a sand dune via a simple text command, highlighting intuitive editing.
Purpose. Illustrate the flexibility in styling and messaging.
Execution. Text overlays change font, color, and content dynamically, demonstrating creative control.
Purpose. Emphasize multi-platform adaptability and global localization.
Execution. A rapid montage of the same image with different aspect ratios and translated text, showcasing versatility.
Objective
A product that automatically generates short-form video ads from long-form content, allowing users to swap out product shots and adjust text overlays for different platforms.How the recipe adapts
The LayerShatterReveal primitive would be used to show a long-form video breaking down into key scenes or elements. InPlaceContentMorph would demonstrate swapping a product shot within a generated ad. DynamicGridMontage would display the same ad automatically reformatted and localized for TikTok, Instagram Reels, and YouTube Shorts, each with different text overlays.Scene durations (5 scenes, 32s)
Key Remotion components
To ensure that artificial general intelligence benefits all of humanity

Gopuff introduces Go, an AI shopping assistant built with SpaceXAI: say what you need and the order is placed.

Iliad (Troy) trailer made by Grok Imagine 1.5, which was just released