Fear in Ratzenstadt

Overview

An AI-generated historical folk horror short created end-to-end through a multi-model, human-in-the-loop workflow.

I built the film from scratch: research, story, screenplay, storyboard, visual development, audio, video generation, editing, and final delivery.

The production combined Veo 3.1, Wan 2.6, Midjourney, ChatGPT, ElevenLabs, Photoshop, and Premiere Pro, using each tool for the parts of the workflow where it performed best.

Every stage was reviewed and directed by me. Research, scripts, compositions, performances, camera movement, lighting, continuity, historical detail, sound, and pacing were iterated until they worked as part of one coherent film. Failed generations were rejected, reworked, or regenerated.

The final piece was selected, edited, paced, and assembled by a me, with AI functioning as a set of production tools rather than an autonomous filmmaker.

The project reflects how I work with emerging AI technology: test models quickly, understand their strengths and limitations, combine them into a practical production pipeline, and turn an original idea into a finished cinematic piece.