Summary
This tutorial walks through a complete AI video production workflow built entirely inside Higsfield, covering everything from prompt writing to final generation in about 14 minutes of screen time. The creator demonstrates using Seedance 2.5 for video generation with synchronized audio, showing how timestamped prompts (describing what happens in the first few seconds, middle, and end of a shot) produce more coherent motion than generic one-line prompts.
The video then moves into image-first workflows using Nano Banana Pro to generate a starting frame before animating it, which solves the consistency problem of getting the same face or character across multiple generations. The creator builds a personal “character sheet” from a selfie and uses it to generate multiple video scenes, including a tea-pouring shot and a basketball action sequence, while explaining how specific photography terminology (like naming an 85mm lens or “sports photojournalism”) produces more realistic results than vague requests.
Throughout, the video emphasizes how sound design and precise timing in prompts matter as much as visual quality, and shows practical techniques anyone can apply for building AI-generated video content with a consistent character across scenes without cycling through multiple separate tools.
📺 Source: Youri van Hofwegen · Published September 28, 2026
🏷️ Format: Tutorial Demo
