Thumbnail

This page shows you how to make a YouTube thumbnail with your face, your title and your words, and how to check it at phone size.

What it does

Thumbnail makes click worthy thumbnails for YouTube. You tell it what the video is about, pick a category and a template, and add your face or a product photo. It makes up to three different concepts at once.

The words on the thumbnail are drawn by Thumbnail on top of the picture, not by the image model. So Hindi and other Indian scripts spell right, and you can change a word later without making a new picture.

When to use it

• You upload videos to YouTube and want a thumbnail that fits your category.
• You want your own face in the thumbnail, or a ready presenter if you do not have a photo.
• You want to see two or three ideas and pick the best one.
• You need a Shorts cover or a square podcast cover from the same idea.

For a poster, a festival post or a menu, use Graphic Design. For a plain picture with no thumbnail rules, use Create.

Step by step

1. In the menu, open Design Studio, then Thumbnail. You see a prompt box that asks what the video is about.
2. Type what the video is about, for example "I tried a 500 rupee phone for 7 days". Hindi, Hinglish or English is fine. A line under the box, Auto pick from your words, shows the category it guessed. If it is not sure, it asks "Is this" a category, and you press the right one.
3. Press Direction. The Direction sheet opens for Thumbnail.
4. Under YouTube category, check the card is right. Under Template, pick a layout, or press Auto pick.
5. Under Video and text, type your Video title, then the words for the thumbnail. Keep to a few big words.
6. Under Essentials, open Person and images. Pick the source Upload photos, then add your face in the Your face slot. The first time, read Your responsibility and press I agree, add the photo.
7. Under More options, choose how many Concepts you want: 1, 2 or 3. It starts at 3.
8. Press Apply direction.
9. Check the number on Generate, then press it. Your thumbnails appear under Your creations and in your Library.

Screenshot coming soon

If the template needs a person and you added no photo, the tool asks before it makes anything. Press Add my photo or Use a presenter. If the template needs a product and you added none, press Add photo or Continue without.

Screenshot coming soon

Designed frames

Some accounts also see a Designed row above the templates. A designed frame gives a finished thumbnail: very big words behind your head, a progress bar, glowing circles, a curved arrow or two stat cards, a dark studio look and a thin bright border. There are six frames: Tool Orbit, Giant Number, Income Banner, Glow Ladder, Low To High and Two Truths, in the same categories as the templates.

1. In Direction, under Template, pick a card in the Designed row.
2. Under Video and text, type your own words and numbers in the frame's fields, for example Big figure ₹5,00,000 and Bar percent 72. The grey words are only examples and are never drawn.
3. Pick the person or the product, then press Apply direction and Generate.

The picture model makes only the photo. Thumbnail finds the face, cuts the person out on our own server and draws every word, number and shape itself, so the words and Rupee figures come out exactly as you typed them, in any script. The big words can sit behind your head. When the person cannot be cut out, or the frame shows a product, the words and shapes go on the side away from the face, and the photo keeps its own light. Before you see the result, the Thumbnail Lock checks it: every word drawn as typed, big enough at phone size, no number hidden by the head, nothing of ours over the face, nothing cut off at an edge, words that stand out, nothing in the corner where YouTube shows the video length, and a face big enough. If something fails, it moves the words or the person and draws again. What it cannot fix shows under Checks on the result.

Edit text on a designed result shows the frame's own fields. A new number is drawn and checked again for free.

Change the words, or check the result

1. Press a result to open it.
2. Press Edit text, change the words, and press Redraw text. The picture stays the same and the edit is free.
3. Press Previews to see it in the feed, in search and at a glance.

Every option explained

The page

The Direction sheet

The person sources are Upload photos, My trained face, Kritiverse model and Describe a person. A trained face comes from Models. A Kritiverse model is a ready AI presenter, not a real person.

In the live preview you can also change the picture shape: 16:9, Shorts, Podcast or Pair. Pair makes a 16:9 and a 9:16 from one direction and counts as two concepts.

On a result

Who can use it

Free plan: not included. Pro plan: included.

What it costs

The number on Generate is the price of all the concepts you chose. It shows before you press it. Edit text and Redraw text cost nothing. The Master file and Make variations are paid and show their price on the button.

Every thumbnail is final and paid once it is made. If a job fails, you see "Nothing was made" and the credits come back.

Tips

• Use 3 to 6 words, fewer for Shorts. The sheet shows a count such as "3 / 5 words" and warns when you go over.
• Do not repeat your video title. The check says "The hook repeats the title. Say something the title does not."
• Add your face photo with a clear front view. A face that is too small at phone size is flagged after the run.
• In the live preview, switch on Squint and check the 160 px size. If it is not clear there, it will not be clear on a phone.
• Keep the bottom right corner clear. The video length covers it.
• Pick 3 concepts when you are not sure. Use Make variations on the one you like best.
• You can see a thumbnail on the YouTube home and on mobile in the separate Thumbnail preview page.

Common problems

"Pick a thumbnail category first." Open Direction, choose a YouTube category, and apply.

"This template needs a person. Add your face photo, or pick another person." Add your photo in the Your face slot, or choose a Kritiverse model.

"Confirm your responsibility for the face photo first." Add the photo again and press I agree, add the photo.

"That trained face is not ready. Pick another one." Pick another face under Person and images, or finish training it in Models.

"Write an idea, add an image or pick what you are making." The prompt box is empty. Type what the video is about.

"These words make a claim we cannot show, such as guaranteed returns or a cure. Change the words. No credits were used." Rewrite the claim, then try again.

"This goes against our content rules, so it cannot be made. No credits were used." Change the words or the picture.

"Text could not be drawn." Press Try again. It draws the saved words again and costs nothing.