🎬 How We Made a 3-Minute AI Video for $400 Using Google Veo3

We created a cinematic launch video for Origami using only AI. It looks nearly human-level realistic — and took just 2 days and $400 in credits. Here's exactly how we did it, step by step:

🔧 Tools We Used:

• Google Veo3 – for high-fidelity AI video generation
• ChatGPT + Claude – for scripting and scene ideation
• CapCut – for editing, sequencing, transitions, and music

🪜 Step-by-Step Breakdown:

1. Scriptwriting (1–2 hrs)

We used ChatGPT and Claude to write a tight, cinematic 3-minute script.

• Broke it down into 15–20 shots for clear visual mapping
• Made sure each shot had a purpose, mood, or transition
• Plan for each shot to be 1-5 seconds, it’s difficult to string together longer shots than this

2. Scene Prompting (~2 hrs)

For each shot, we crafted detailed Veo3 prompts.

Example: “Close-up of a glowing neural network, pulsing light, futuristic, macro lens”

• Kept track of prompt tweaks for each iteration

Here’s an example of one of ours: