Filmmaker builds ComfyUI workflow that auto-generates all shots for his short films
Paranormal-Dream · reddit · 2026-10-12
A filmmaker with a traditional production background shares a ComfyUI workflow that turns a short film concept and voiceover script into visual shot options. An LLM (ChatGPT) trained on his own examples of translating narration into scenes and thematic analogies — plus a style LoRA for aesthetic control — generates the shots, which he then curates. Pre-production and post-production stay fully manual; only 'production' is automated. Built on Mac using Floyo as the ComfyUI environment, developed during a 13-week Advanced ComfyUI program at Lighthouse AI Academy, and used to make the short film THE MIRROR.
More from Multimodal
- ByteDance world model demoed in five use cases, from film camerawork to robot training — oran_ge · 2026-10-12
- Unlimited OCR: 3B open-source model reads 100-page PDFs in one pass, but scores just 46.2% on ParseBench — solyarisoftware · 2026-10-12
- Flux Metal Jacket V15: One 257-Node ComfyUI Workflow Bundling 12 Image Models — CableNo3994 · 2026-10-12
- Photo-real Chinook monster truck: locking first frames with Magnific One — techhalla · 2026-10-12
- Nostalgia for the Deforum era: one AI video took 4 months and 100+ versions — ART-ficial-Ignorance · 2026-10-12
- AI-generated Korean mech blockbuster poster looks straight out of a theater lobby — Foreign-Original124 · 2026-10-12