PROMPT GUIDE

How to make YouTube thumbnails with AI (16:9 and space for text)

A thumbnail is a picture people glance at on a small phone screen before deciding whether to click. When making one with AI, prioritize a clear subject, readable text and a hint of what the video is about over a pretty picture. This guide covers framing and prompts that suit YouTube thumbnails, and the steps for adding text.

Last updated: 2026-10-22 approx. 3 min read

日本語版

The short answer

When making a YouTube thumbnail with AI, generate a wide 16:9 image, place the person or main subject on the left or right, and leave empty space on the other side for text. Writing on the right third of the frame, empty space on the left in the prompt creates that space. Do not let the AI draw the text; add it later, big and bold, for a readable thumbnail.

How do you make a YouTube thumbnail with AI?

Use a wide 16:9 image, put the subject on one side and keep the other side empty for text, which you add later.

Thumbnail size and ratio rules are set by YouTube and can change, so check the official help for the latest recommendations. A wide 16:9 ratio is commonly used. When generating with AI, start with a ratio close to 16:9 (for example 1344x768) so important parts are not cut off when you crop later.

Words that suit thumbnails

Tag	Meaning	Note
`on the right third of the frame`	Person in the right third of the frame	Leaves the left side for text
`empty space on the left`	Empty space on the left	Where the text goes
`close-up`	Close-up	The face reads even on a small screen Details
`surprised`	Surprised face	An eye-catching expression for thumbnails Details
`pointing`	Pointing	Draws the eye to the text or an object
`simple background`	Simple background	Makes text easy to read Details
`vivid colors`	Vivid colors	Stands out even when small Details

Wide thumbnail-style illustration with a girl on the right and empty space on the left — 右に人物・左に余白

Anime girl pointing at empty space on the left, a thumbnail layout example — 右に人物・左に余白

Illustration usage rules

What are the steps to write a thumbnail prompt?

Steps to make a thumbnail with AI

Decide the thumbnail text that sums up the video first (a few short words are easiest to read)
Decide whether the text goes left or right, and place the subject on the opposite side
Choose one expression (surprise, smile, worried face) that matches the text
Make several wide images close to 16:9 and pick one whose face reads on a small phone screen
Add the text big and bold in an image editor, with an outline or banner so it stands out from the background
Finally check the text is readable when shrunk, and save with light compression

1girl, solo, surprised, open mouth, wide eyes, hands on own cheeks, upper body, on the right third of the frame, empty space on the left, simple pastel background, vivid colors

an anime girl with a surprised face pointing to the left, upper body, standing on the right side of a wide frame, the left half is a plain bright yellow background with nothing on it, bold clean colors

Which expressions and framing catch the eye?

Big expressions and larger faces catch the eye. Matching the expression to the video's content also avoids disappointing viewers who click.

Video content	Expression and pose	Example words
Surprising results, experiments	Surprised face, open mouth	surprised, open mouth, hands on own cheeks
How-to, explanations	Pointing, smiling	smile, pointing, looking at viewer
Mistakes, warnings	Worried face, sweat drop	troubled, sweatdrop, frown
Rankings, introductions	Hand sign, energetic smile	grin, v, energetic
Calm work, chatting	Gentle smile	gentle smile, relaxed

Rinka's tips

Make the background behind the text plain or heavily blurred (simple background or depth of field)
If the face is at least a third of the frame height, the expression reads on small screens
Using the same character and layout every time gives your channel a consistent look

To find expression prompts (surprised, worried, smiling): Facial Expression Prompt List (Smile, Blush, Crying) →

What are the common thumbnail mistakes and fixes?

Problem	Cause	Fix
No room for text	The person is in the center	Add on the right third of the frame, empty space on the left
Unrecognizable when shrunk	The person is small or the background is busy	Use close-up or upper body and a simple background
Made it portrait, and it gets cut when made wide	Started with the wrong ratio	Generate wide, close to 16:9, from the start
Meaningless text appears	The word thumbnail makes it imagine text	Add no text and add the text later
Dull colors that get lost	No color instruction	Add vivid colors, high contrast

What should you watch when making thumbnails?

Avoid thumbnails that misrepresent the video, and images that imitate other people's works or celebrities. Check YouTube's official policies for thumbnail rules.

Clickbait thumbnails that differ greatly from the content lower viewers' trust
Do not use images that imitate real people or other people's characters
Check the generation service's terms to see whether images can be used in monetized videos
File size limits are set by YouTube, so check the official help and compress if needed

How to add clean text to thumbnails: How to Put Text in AI Images (Prompts That Avoid Garbled Text, and Adding It Later) →

Wide sizes close to 16:9 (such as 1344x768): SDXL Resolutions (Recommended Sizes) and How to Pick a Ratio →

To make a header image (eye-catch) for a blog post: How to make blog featured images with AI (OGP ratio and margins) →

FAQ

Frequently Asked Questions

What ratio should a thumbnail be?

A wide 16:9 is common. Generating close to 16:9 from the start means important parts are not cut when cropping. Recommended sizes can change, so check YouTube's official help.

Can AI add the thumbnail text too?

Text breaks easily, so adding it later in an image editor is recommended. Write no text in the prompt and let the AI create only the empty space for it.

How do I place the person on one side?

Write both the position and the side to keep empty, such as on the right third of the frame, empty space on the left. A simple background makes the empty space easier to use for text.

Which expressions suit thumbnails?

Clear expressions such as surprise, a smile or a worried face catch the eye. Choosing one that matches the video's content also prevents disappointment.