Step-by-step production pipeline for faceless AI room-restoration YouTube Shorts. One step per message: premise → named images (Nano Banana) → videos from those images (Vidu Q2) → voiceover (ElevenLabs v3) → CapCut placement → YouTube settings and publish. Use when the user says "new short", "restoration short", "episode", "next", "run the pipeline", or asks for hooks, scene lists, image/video prompts, VO scripts, titles, or tags for the restoration channel.
