Free · Five steps · No signup
The AI-UGC system.
Five steps from a blank page to a clip somebody actually sees. Four of them are craft and you can do all of it yourself with what's here. The fifth is infrastructure, which is the part that stops most people.
The script
Structure, sentence rules, and the writing prompt.
Read›The image prompt
The six-part skeleton, per-format values, real renders.
Read›The clip and the edit
What to append to every i2v prompt, and the edit spec.
Read›Checking it
The review prompt and what reads as generated.
Read›Getting it seen
Why a good clip on one account still dies, and what it costs to fix.
Read›Prompts, not theory
A six-part image skeleton, the two clauses that belong on every video prompt, and a review prompt. Copy them.
Values off real winners
Camera height, framing, eyeline and set for each format, read frame by frame off clips that performed, with links so you can check.
Our own renders
Clips we made, including one that ignored its own prompt, so you can see the difference rather than take our word.
Where it stops being craft
Step five is honest about what you cannot prompt your way past, and what running it yourself actually costs.
The shortcut
Skip the reading. Get the system prompt.
One block you paste into Claude or ChatGPT once. After that you hand it a product and a format, and it writes the image prompt, the video clauses and the script to every rule in the first four.