I Recreated a Spider-Man Scene at Home Using AI (Full Tutorial)
Demonstrates how filmmakers without 3D or VFX expertise can recreate complex cinematic stunts using keyframe spatial storyboards and low-res draft testing.
What this lesson covers
Kaye Creatives provides a step-by-step tutorial on recreating an iconic animated scene from Spider-Man: Into the Spider-Verse using text-to-video generative AI tools. The process covers character asset creation with Seedream 5.0 Pro, keyframe mapping via Claude, and low-cost motion rendering with Seedance models in Higgsfield AI.
Key takeaways from the creator
AI-extracted notes, not independently verified product claims. Timestamp links let you check each point in the original video.
- 01:33 ↗
Construct 360-degree character sheets as reusable assets so AI video models maintain consistent outfit and facial features across shots.
- 02:37 ↗
Use Seedream 5.0 Pro for character sheet generation due to its high graphic contrast and sharp line crispness.
- 05:05 ↗
Extract exact keyframe screenshots from source films and feed them to Claude alongside character sheets to map out spatial storyboards.
- 07:08 ↗
Run first-draft video generations at low resolution (480p) with a batch size of 1 to test motion without burning monthly credits.
- 08:04 ↗
Treat AI video generation as an iterative trial-and-error process by feeding specific failure details back into Claude for revised prompts.
Workflow outlined in the video
- Gather mood and streetwear references on Pinterest to style a custom character concept.
- Prompt Claude with selfies and references to generate a full character sheet, then render batches in Cinema Studio using Seedream 5.0 Pro.
- Capture keyframe screenshots of a scene's spatial sequence to establish environmental layout and action beats.
- Input keyframes and character assets into Claude to formulate frame-by-frame text-to-video prompts with specific camera movements.
- Render initial video drafts in Higgsfield at 480p using Seedance models to evaluate action without incurring high credit costs.
- Refine failed motion elements by describing errors back to Claude before generating high-resolution final outputs.
Before you use this workflow
These notes describe the source video at its publication date. Model access, pricing, connectors and interfaces may have changed. Check the original source and the provider’s current documentation before installing an add-on, connecting an account or spending credits.
ReelStack has not independently tested this workflow. Preview one representative shot and check motion, continuity and output quality before applying it to a full production. No result, cost saving or model capability is guaranteed.
Explore the tools
How to interpret the numbers
116,779 views captured 27 September 2026. ReelStack recommendation score: 44/100. Formula: radar-v5. This is ReelStack’s calculation, not a YouTube rating or a measure of factual accuracy.
Show the calculation
- 65% performance: 1.01× views vs 3 other channel videos observed at a similar age. View-evidence factor 99% (views / (views + 1,000))
- 25% freshness: 16/100 with a 14-day half-life
- Momentum pending: collecting daily snapshots; its 20% weight goes to observed performance, not free points
- 10% engagement: 73/100; likes + 4× comments, smoothed with a 500-view neutral prior