Writing an AI text prompt
A prompt will generate short, concise, and compelling ad copy, so it has a limitation in characters’ number to perform better on short texts like headlines and descriptions.
It will always give you 5 text options when you generate a copy based on a prompt, regardless of asking for a different number of options.
Here’s how to write a good prompt so you get the best results from it.
It’s important to offer context - tell the AI what you want to generate in detail, as the more information it has, the more likely it is to give you more accurate results. It helps when you mention what type of text you want (headline, ad description).
Example:
- Good prompt: write a headline for an online ad about an exotic vacation in a luxury hotel in [country]
- Basic prompt: special offer at a luxury hotel
You can combine your prompt with an example text based on which the AI will give you options. That means you should write the text on your ad and select “Based on selected layer” together with your prompt when you generate with the AI.
Writing an AI Image and Video prompt
To help you write better prompts, we've introduced the Enhance prompt. This functionality automatically enriches your prompts to improve the images and videos generated by the AI. You can find it in the prompt box.
Tips for crafting effective prompts
- Be detailed. In your prompt, provide as much detail as possible to avoid ambiguity. The more specific you are, the better the AI can understand your vision.
- Avoid ambiguous language, such as vague terms like make or generate. Instead, focus on descriptive language that depicts your desired outcome.
- Balance complexity. Strive for a prompt detailed enough to guide the AI accurately but not overly complex. Short prompts may result in unintended outcomes, while overly elaborate prompts can be challenging to tweak afterward.
- Combine keywords. You can combine multiple keywords from the same category within your prompt, but avoid overwhelming the AI with too many similar specifications.
- Tweak as needed. The AI's interpretation may require adjustments. Be prepared to iterate on your prompt and fine-tune the result based on the generated image or video.
Understanding the structure of a prompt
Before diving into the specifics, let's break down the components of an effective AI image and video prompt.
Subject - What you want to see in the image or video. You can add names of famous people, animals, plants, objects, etc.
Example: Majestic dragon perched atop a snow-capped mountain.
Medium - The material used to make the artwork.
Example: oil painting, portrait, digital painting, photography, drawing, concept art.
Style - Artistic style¹ of the image or video
Example: hyperrealistic, pop-art, art nouveau, modernist, fantasy.
Framing - Here, you can name the composing view.
Example: close-up, full body, upper body.
Camera Angle - Name the specific position of the camera.
Example: from above, high angle, from below, low angle, from behind.
Artist - Type the name of an artist to mimic their style.
Example: John Singer Sargent, Alphonse Mucha, John Collier.
Website - You can even type in specific niche graphic websites.
Example: artstation, deviant art, pixabay.
Resolution - How sharp and detailed the image or video should be.
Example: highly detailed, sharp focus, 8K, unreal engine.
Color - Define the overall color of the image or video. You can also specify parts of the image or video together with a color
Example: vintage, iridescent gold, silver, blue background.
Lighting - Specify what type of light to be in the image or video or the absence of it.
Example: rim lighting, cinematic lighting, crepuscular rays, dark.
💡It's important to note that you don't need to use all the components laid out in our example; you have the flexibility to experiment, mix, and match so you get the result that works for you.
Note¹: If you're selecting one of the styles available in The Brief (Realistic, Digital art, 3D model, Isometric, Pixel art, Anime, Vaporwave), you don't need to type in the style in the prompt.