1. Workflow Architecture
- Trigger: Webhook Trigger receiving a single text parameter
topic(e.g., "The History of Rome's First Emperor"). - AI Director: The AI Storyboard Director (OpenAI) node generates a high-signal, 3-scene detailed storyboard package (comprising voiceover scripts, visual prompt tags optimized for Flux, and motion vectors for Runway).
- Loop Routing: The Split Scenes node runs the rendering pipeline for all scenes in parallel.
- Audio Generation: The ElevenLabs Voiceover node converts scene script nodes into realistic, premium human voice narration audio.
- Image Synthesis: The Flux Image Generator (HTTP Request via Replicate API) creates 16:9 cinematic aspect-ratio high-fidelity base frames.
- Flux Verification Loop: The Poll Flux Status and Flux Ready? (If + Wait) nodes continuously check Replicate's polling URL until rendering is completed.
- Video Animation: The Runway Video Gen node maps the synthesized Flux image and the AI-generated motion vectors, transforming static frames into 4-second cinematic video scenes.
- Runway Verification Loop: The Poll Runway Status and Runway Ready? (If + Wait) nodes verify completion of the video rendering task.
- Data Assembly: The Aggregate Storyboard node merges audio elements, video scenes, and scripts back into a unified array.
- Airtable Command Center: The Save to Airtable node pushes the final aggregated scene assets (script text, rendering outputs, audio downloads) to a collaborative production table.
- Instant Notification: Parallel Slack and Telegram nodes dispatch a high-priority message containing executive storyboard notes and immediate workspace links.
2. Setup Instructions
- Replicate API (Flux): Acquire a developer token from Replicate. Insert it under the authorization header in the Flux Image Generator and Poll Flux Status nodes.
- Runway ML API: Register on Runway ML's developer console, generate a key, and update the headers in Runway Video Gen and Poll Runway Status nodes.
- ElevenLabs API: Retrieve your XI API Key and target Voice ID, inserting them into the text-to-speech HTTP Request node.
- Airtable Setup: Construct an Airtable base with table name
VideoDashboardand columns for:Topic,Scene 1 Video,Scene 1 Audio,Scene 2 Video,Scene 2 Audio,Scene 3 Video, andScene 3 Audio. Copy your Base ID and Table name into the Airtable node.
