Best Practices

Be Specific

Vague prompts produce unpredictable results. Include concrete details:

Instead of	Say
"Make a short video"	"Create a 10-second video"
"Add some text"	"Add 'Welcome' in white, centered on screen"
"Make it look nice"	"Use a fade-in animation on a blue background"

Describe Timing

Specify when elements appear and how long they last:

Show 'Introduction' for 3 seconds, then transition to 'Chapter 1' for 5 seconds, then 'The End' for 2 seconds

Specify Layout

Describe where elements should appear:

Place the logo in the bottom-right corner with the title centered above it

Request Animations

Name the effect you want:

Add a typewriter animation to the title text

Layer Elements Correctly

When combining elements, specify what's in front:

Show 'Breaking News' text over a city timelapse background

The AI will place text in the foreground and video in the background.

Include links to assets

LLMs have the tendency to hallucinate URLs that don't point anywhere. Provide direct publicly accessible URLs for all assets, including fonts.

Use the Arimo font for the header (https://shotstack-assets.s3.amazonaws.com/fonts/Arimo.ttf)

Iterate

Start simple, then refine:

1. "Create a 10-second video with 'Hello World' centered"
2. "Make the text white on a dark blue background"
3. "Add a fade-in animation to the text"
4. "Add gentle background music"

Use Templates for Variations

For multiple similar videos, create a template first:

Create a template called 'greeting' with a NAME placeholder, then render it with NAME set to 'Alice'

See Managing Templates for details.