Search

Educational AI Daily

Artificial intelligence, as told by three teaching communities

Multimedia 1 November

AI audiovisual

An AI-made horror web series documents its workflow step by step

“Never Sleep Again”, seven episodes released between 25 and 31 October, explains on its website how it combines Blender, Claude, a video generator and DaVinci Resolve.

“Never Sleep Again” is a seven-episode horror web series created by Raül Torres with AI tools under human direction. Six episodes run between 29 and 42 seconds, and the finale, released on Halloween night, lasts 2 minutes and 8 seconds. They were published daily from 25 to 31 October on YouTube, Instagram, LinkedIn, TikTok and X.

The most useful part for teachers interested in audiovisual work is that the series’ page details the method. The house where the story takes place is modelled in 3D with Blender, so every shot is filmed from a camera placed in a coherent space. For each camera movement, a grey reference clip with no visual style is generated and handed to a video generator (Seedance 2.5, in Dreamina) along with scene images and instructions. Claude, connected to Blender and Dreamina, helps prepare the references and write those instructions. Each result is checked against the plan and regenerated with tighter references if necessary.

Editing and colour correction are done in DaVinci Resolve Studio, and Fusion is used to add, as separate layers, the phone’s video-player interface and a night-camera finish. The author told the group that he has completely changed the way he works with Claude’s connectors (MCP), which lets him handle Blender and DaVinci Resolve at a level he could not reach before.

Read the original conversation in EduMedia-IAG

Checked in

Chalk and Algorithm's comic strip 1 November

Halloween nightI've made you a horror series.Watch out for the monster.Let's see if it's scary.
A dark house. A doorcreaks. And suddenly...
“Hello, I'm the monster.How can I help youtoday?”

Comments

Commenting requires a free GitHub account. Comments are public and stored on GitHub.

Back to the front page