Summary
Youri van Hofwegen walks through an end-to-end workflow for producing History Channel-style documentary videos entirely with free AI tools, from script to finished video. The pipeline begins with a master prompt pasted into Claude, which outputs a full voiceover script, scene-by-scene breakdown, an image prompt per scene, a matching motion prompt, and even music and thumbnail suggestions — all from a single interaction.
For narration, the tutorial uses Google AI Studio’s speech-and-music playground with the Gemini 3.1 Flash TDS preview model and the ‘Enceladus’ voice set to empathetic delivery. A key tip is filling in the scene context and sample context fields — describing the studio setting and delivery style — which the video argues is the difference between a flat AI read and a performance that sounds authentically documentary-style. Images are generated in Google Flow using the Nano Banana Pro model at 16:9 ratio for a cinematic film look, with Claude’s self-contained prompts handling consistency across scenes and keeping real historical figures’ faces obscured to avoid synthetic likenesses. Motion is added through Meta AI using a free Chrome extension called Auto Meta that batch-processes all scene images in one pass.
The tutorial is honest about the ceiling of free tools — face consistency between shots and overall cinematic quality lag behind paid options — and positions the workflow as a starting point for creators who want to experiment before committing to paid AI video platforms.
📺 Source: Youri van Hofwegen · Published July 08, 2026
🏷️ Format: Tutorial Demo
