NodeBB

Tried a variation on this one: same camera, same beats, but a house instead of a tower. The chain held and the framing stayed clean all the way through. If you run it, post what you get.

A prompt is not a wish. It is a specification. Four parts, and if one is missing, the output wanders.

Method. The method picks the structure. A time-lapse is built differently from a product shot, and a product shot is built differently from a short drama. Pick the method first, and the rest of the prompt follows from it. The free library holds fifteen methods, each one a complete structure for one kind of output.

Beats. Beats carry the motion. A prompt that says "a city grows" gives you a blur. A prompt that says "empty lot, foundations, steel frame, glass, lights on" gives you a sequence. Write the beats down, one per shot, before you generate anything.

Settings. Settings fix the output. Model, aspect ratio, duration per shot. These are not decoration; they are part of the specification. The same words on a different model or a different ratio are a different prompt.

Keyframe. The keyframe locks identity. Camera, subject, light, locked in the first frame and held through every shot. Without it, the subject changes face halfway through and the sequence falls apart.

Then iterate. Change one thing at a time. A new subject, same camera. A new camera, same subject. Keep notes on what worked, and post the working version here in Prompt Sharing, in the format from the pinned post.

That is the whole craft. Method, beats, settings, keyframe. Everything else is practice.

A model drifts when the words are loose. So lock the words first, and lock the frame first.

The chain we use:

1. Idea: what the shot is, in one sentence.
2. Keyframe: the first frame, locked. Camera, subject, light. Everything else is built off this frame.
3. Beats: the motion, broken into shots. One beat, one generation.
4. One generation per shot: five seconds at a time, no more. Short generations stay clean.
5. Continuity: same camera, same subject, same light from shot to shot.
6. Assembly: shots in order, then voice and sound last.

The words are the method. Write them down verbatim before you generate, and keep them the same across every shot in the chain. Change one thing at a time when you iterate, or you will not know what worked.

This is how the 15 methods in the free library are built. Each one is a locked version of this chain, tuned for one kind of output.

This is the home of prompting. A forum, a chat, and a bulletin board in one place, built for people who make things with AI video.

Everything here is free and open. Prompts, techniques, and methods are shared freely. Nobody paywalls a prompt here. That is the whole point: the free library draws people in, and this community keeps them here, trading what works.

What lives here:

• Prompt Sharing: post your prompts, and post what they made.
• Techniques & Methods: how prompting actually works, method by method.
• Blogs: longer guides and deep dives.
• General Discussion: everything else.

Start by reading "How to post a prompt here" in General Discussion. It shows the format. Then post your first prompt. That is how the cycle starts.