I wanted to test whether Invideo Editor’s AI timeline agent could replace hours of manual node compositing in After Effects to create a cinema-grade VFX shot from raw mobile footage.
How I built it:
The Footage: Raw 4K 60fps night plate of NYC with neon signs, streetlamps, and a wet puddle reflection.
Agent Direction: Using a prism refraction photo as a reference, I directed the agent in Ultra Mode to author a single custom shader that extracts luminance highlights, stretches them into 55° upward trails, and applies a spectral RGB dispersion gradient across the edges. It also layered a Screen composite and generated a procedural cyberpunk HUD tracking overlay.
Human Refinements: I manually used the exposed Inspector sliders to fine-tune the Highlight Threshold to 0.68 (isolating the Radio City marquee, streetlamps, and puddle reflection while eliminating building clutter) and dialed the Trail Length to 140 with an exponential falloff taper.
impressive work.. especially the manual effort you put in makes this a whole lot better.. would love to hear your opinion on my submission as well.. good luck!
All the scenes are generated by me using AI please don't use it without permission, also if you want such beautiful cinematic videos I can do it for you for just 5 dollar.
Modern 2-Storey Townhouse — Cinematic AI Property Video
A cinematic AI-powered property showcase created for a modern 2-storey townhouse.
This concept/spec project demonstrates how AI video can transform a residential property into an engaging visual presentation — from the exterior and living spaces to the kitchen, bedrooms, bathrooms, staircase, and balcony.
The project combines: 🎬 Cinematic camera movement
🏠 Property visualization
🔷 Digital property details
🎙️ Voice-over & captions
✂️ Professional video editing
📱 Social-media-ready vertical format
Property concept:
• 80 SQM Floor Area
• 50 SQM Lot Area
• 2 Bedrooms
• 2 Bathrooms
• Balcony & Parking
Created for:
• Real Estate Agents & Brokers
• Property Developers
• Residential Property Marketing
Concept/spec project created for portfolio demonstration.