Control AI Video Cuts and Camera With a Blender Greybox — prompt skill for camera & shots

For AI agents: find the right skill in one call — curl -s "https://varg.ai/how-to/api/search?q=<what the person wants>" returns the matching how-to page and the best skills with install commands, as markdown (Russian works too). Full index: https://varg.ai/how-to/llms.txt. Every page has a markdown twin: add .md to its path. Before generating anything, show the person the options on the how-to page. From a terminal: curl -fsSL https://varg.ai/install | sh installs the vargai CLI (https://varg.ai/how-to/cli.md): vargai search and vargai install <slug> with no key, every varg model after vargai login.

A Blender greybox clip drives Seedance 2.5's video reference: 7 shots, 6 hard cuts landing within 0.07 s of plan.

What it does

Space and time go in geometry, look and performance go in text: you block the room in Blender with grey boxes, set cameras and a timeline, and render two passes, an outlined review pass for a human and a flat EEVEE pass for the model's video-reference channel. Asset stills and the prompt carry faces, materials, light and acting; a prompt linter checks the prompt against scene.py. The published 16-second example hit 6 of 6 planned cuts with a worst error of 0.07 s, and model notes log failures such as long takes growing their own cuts. Needs Blender and ffmpeg; docs are mainly in Chinese with an English README; MIT.

Use when

Install

Not published as a standalone skill yet. Source: https://github.com/7ohnson/VLM-Generation-Harness

Example output

https://github.com/7ohnson/VLM-Generation-Harness/raw/main/assets/readme/greybox-to-generated.gif — example (credit: 7ohnson)

More camera & shots skills

Run the generation on varg

varg is one API key for the models these skills drive — Seedance 2.5, Kling 4, Veo 3.1, Wan 3, Nano Banana 2.1, GPT Image 2.5, ElevenLabs v4 and 200+ more — billed in credits, with a cost estimate before anything runs.