Opus 5.5 Makes INSANE Simulation Videos!
Relevant for AI filmmakers interested in programmatic generation, interactive 3D simulations, and procedural documentary creation using advanced coding LLMs paired with voice APIs.
What this lesson covers
Adrian explores programmatic video and simulation generation using Claude and Opus 5.5 across documentary, architectural, and commercial use cases. He explains how setting code effort levels and integrating external API keys like ElevenLabs produces complete scripted animations and voiceover.
Key takeaways from the creator
AI-extracted notes, not independently verified product claims. Timestamp links let you check each point in the original video.
- 01:25 ↗
Connecting an ElevenLabs API key directly into the workspace folder significantly elevates the character narration voice quality.
- 02:16 ↗
Increasing the coding effort slider to Max or Ultra Code produces complex, multi-layered simulations like architectural builds, though generation can take up to 45 minutes.
- 04:40 ↗
A single prompt with high-effort code parameters can generate both an interactive 3D historical simulation and an authored documentary video with camera angles.
- 06:12 ↗
Because programmatic simulations are code-based, minor director tweaks (such as camera speed or color changes) can be re-recorded without rerendering the entire asset from scratch.
- 06:27 ↗
Giving the model reference video links or descriptions enables it to reverse-engineer narrative structures and replicate formats for commercial ad spots.
Workflow outlined in the video
- Store your ElevenLabs and image generation API keys in a scoped folder accessible to the model.
- Provide a reference image or storyboard layout to establish character design and video aspect ratios (16:9 or 9:16).
- Adjust the coding effort parameter to Max or Ultra Code when generating complex timeline simulations or data-driven environments.
- Prompt the model with specific instructions for multiple camera angles, narration scripts, and caption styling.
- Iteratively refine specific visual elements (such as camera pan speeds, colors, or added assets) using programmatic updates rather than re-generating from scratch.
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
4,379 views captured 1 October 2026. ReelStack recommendation score: 57/100. Formula: radar-v5. This is ReelStack’s calculation, not a YouTube rating or a measure of factual accuracy.
Show the calculation
- 65% performance: 1.61× current views vs the channel's recent median; same-age history is not yet available. View-evidence factor 81% (views / (views + 1,000))
- 25% freshness: 85/100 with a 14-day half-life
- Momentum pending: collecting daily snapshots; its 20% weight goes to observed performance, not free points
- 10% engagement: 27/100; likes + 4× comments, smoothed with a 500-view neutral prior