PROMPT GUIDE

Write AI video effect prompts with clear timing and motion

You do not need professional VFX terminology. Identify the subject, the moment the effect should begin, the visual change you want, and the overall mood. That is enough to give the production process a clear direction.
Create a video
Describe visible footage
Name the trigger moment
Keep one visual direction
01

A practical prompt structure

Build the direction from four parts: subject and action + trigger moment + visual change + pace and mood. Start with the most important change, then add atmosphere. Complexity does not help when the requested effects compete with one another.
02

People and milestone moments

“As they exchange rings, draw the light toward their hands. Then let small petals rise slowly while the shot remains warm, soft, and restrained.” This identifies the subject, timing, change, and emotional tone.
03

Sports and fast action

“After the athlete leaves the ground, emphasize the motion path. Add a brief burst of speed lines and light before the landing, with a clean, forceful pace that never covers the body.” Fast footage benefits from a precise trigger and a clear visibility constraint.
04

Travel and landscape footage

“As the shot opens from the foreground into the valley, gradually increase the sense of depth. Move distant mist along the ridge, keep the color naturally cool, and pace it like a quiet film opening.” Landscape prompts can focus on space, atmosphere, and color.
05

Pets and playful moments

“When the dog looks back at the camera, add light hand-drawn lines and a brief trail of small stars that follows the head movement. Keep it playful without covering the eyes or expression.” Even a playful prompt should protect the visual information that matters.
06

How long should an effect prompt be?

One to three sentences is usually enough. Gen2049 accepts directions up to 300 characters. Clarity matters more than length: avoid contradictions and state the subject, trigger, and main change. If you leave the field blank, the system creates a direction from the footage itself.