ScareFactory

ScareFactory Workflow Overview Automation Purpose This workflow automates the end-to-end creation of a horror-themed social media video package, specifically designed for platforms like YouTube Shorts and Instagram Reels. Starting from a single story idea stored in Airtable, it coordinates a series of AI agents and services to generate a complete set of assets: an enhanced story script, a full voiceover, a viral video script broken into scenes, and all the corresponding visual images for the video, including a scroll-stopping first frame. The value of this automation is immense, transforming a multi-hour, highly creative manual process into a minutes-long, automated pipeline. It enables the mass production of high-quality, engaging horror content by automating creative writing, voice narration, content strategy, and visual asset generation, allowing a single trigger to yield a nearly-finished video product. Technologies Used - Core Platform: n8n - Data & Project Management: Airtable - AI Text-to-Speech: ElevenLabs - AI Text-to-Image: PiAPI (Flux1-dev), OpenAI (DALL-E) - AI Language Models: Meta Llama 4 Maverick, Perplexity Sonar Reasoning, OpenAI GPT-4o (via OpenRouter) - AI Framework: LangChain.js (n8n Agent nodes) - Cloud Storage: Google Drive AI Integration Analysis This workflow showcases a sophisticated multi-agent AI system, where different AI models are assigned specialized roles, mimicking a creative production team. 1. The Storyteller (Llama 4 Maverick): The first AI agent acts as a creative writer. It receives the initial story and is tasked with improving and expanding it, ensuring the narrative is compelling and meets a desired length. 2. The Video Strategist (Llama 4 Maverick): The second agent functions as a social media content expert. It analyzes the original story and, guided by an extremely detailed prompt about creating viral content, deconstructs it into a short-form video script. It outputs a punchy caption, a series of 5-second video scenes, and a concept for a captivating first frame. 3. The Art Director (Perplexity Sonar Reasoning & GPT-4o): This role is split between two specialist agents.