Prompt

Turn a few sentences into a finished thumbnail.

Prompt makes a thumbnail from a description alone: in WThumb, an AI thumbnail maker for YouTube, you type what the picture should show into one box, with no reference image or video link, and a new design is drawn from your words. Say who is in it, where they are and the exact words to print, pick your own face or Faceless, and press Make.

Open the Studio

2 free credits (1 credit = 1 thumbnail), no card

How it works

From a blank box to a finished picture.

  1. Open Prompt

    The Open the Studio buttons on this page open it already in Prompt; inside the Studio, press Prompt beside Recreate, above the card. The card then holds one box labelled Prompt, with the Persona and Images tiles under it.

  2. Describe the thumbnail

    Write what the picture should show in up to 2,000 characters, and end with the exact words to put on it if it should have any. Then open the Persona tile and pick a persona for your own face, or Faceless for none.

  3. Pick the frame and press Make

    In the bar under the card, choose 16:9, 9:16 or both, and 1 or 2 designs. The button then says how many thumbnails it will make, and it asks you to sign in first if you have not yet.

The Studio's card in Prompt: the box holding the car review description, the Persona tile with the sample persona Demo, an empty Images tile, and under it the bar with Uses 1 credit, 16:9, 1 design and Make
Prompt in the Studio with the car review's description typed in, Demo on the Persona tile and one 16:9 design set.

Examples

Three descriptions and the thumbnails they made.

Each picture below came from one Prompt run at 16:9, with Demo, the sample persona, picked on the Persona tile and the description beside it typed exactly as shown.

“Me” in a description means the person of the persona you picked, and the words after “Words on it:” are the headline that was printed on the picture.

None of the three says how the words should look, so the white letters of WORTH IT? and the yellow on 50 and on 24H were the design's own choices.

The thumbnail the first description made: the sample persona on the left, chin on his hand and looking thoughtful, a white Tesla Model Y on a hilltop road at sunset on the right, and the words worth it?Car review

What was typed

“I drove a Tesla Model Y for a year and ask if it was worth it. Me on the left with a thoughtful look, chin resting on my hand. A white Tesla Model Y parked on an open hilltop road at golden hour on the right. Words on it: WORTH IT?”

The thumbnail the second description made: the sample persona smiling and leaning back with his hands behind his head in a dark office, HubSpot, Notion and Slack app icons around him, and the words 50 clientsAgency clients

What was typed

“How I run 50 client accounts without stress. Me leaning back relaxed in my chair with both hands behind my head, smiling at the camera, in a dark, moody home office with soft warm light. The HubSpot, Notion and Slack logos as big glossy app icons floating beside me. Words on it: 50 CLIENTS”

The thumbnail the third description made: the sample persona holding up a signed client contract and pointing at it, Calendly, Stripe and Google Drive app icons beside him, and the words onboarded in 24hClient onboarding

What was typed

“How my agency onboards a new client in 24 hours. Me holding up a signed client contract and smiling at the camera, in a dark, moody office at night with warm lamp light on my face and a deep, dark background for strong contrast. The Calendly, Stripe and Google Drive logos as big glossy app icons floating beside me. Words on it: ONBOARDED IN 24H”

Three Prompt runs, each picture beside the exact description that made it.

Writing it

What a description that works says.

Begin with what the video is about, in the words you would use to tell a friend. The design is built around the reason a viewer would click, so a first sentence that states that reason gives it the most to work with.

Next, place the person: say where you are in the frame and what you are doing, such as on the left with a hand on your chin, or leaning back with your hands behind your head. Your photo sets the face, while the expression, the posture and the clothes are drawn to suit the scene.

Then give the setting and its light, like a hilltop road at golden hour or a dark office lit by one warm lamp. A few words are enough, as the background is kept calm and softly focused so the person and the headline stand out.

Name any product, app or logo that belongs in the picture and how it should look, such as big glossy app icons; no logo or app icon is added unless your words name it. A laptop or a phone is seen from the side or from behind, with nothing on its screen, unless you say what the screen shows.

Finish with the exact words after a marker such as “Words on it:”, and they are printed as you typed them, so a few short words read best. With no words in the description the thumbnail has none at all, and a line like “you decide” lets the design add one to three words only where the picture needs them.

The description can be written in any language, and so can the words you want printed.

Face and images

Your face, no face, and pictures of your own.

Make waits for an answer on the Persona tile. A persona puts your own face in the design, while Faceless leaves you out: no host or narrator is drawn in your place, and a person appears only where the description asks for one.

As many as 4 images of your own can go in too, such as a product shot, a place or a logo. Add them on the Images tile or paste a screenshot straight into the box, and each one is numbered @1, @2 and on.

Type @ in the description to put an image where it belongs and say what it is for, and one you never place is set wherever the design has room. A person in an attached image is never taken for you.

Frames and cost

Sizes, designs and what a run uses.

Choose 16:9 for a video, 9:16 for a Short, or both; with both, each picture is arranged for its own shape, so the tall one is never a slice of the wide one.

Ask for 2 designs to get two different takes on the same description. Every thumbnail uses 1 credit, so both frames with 2 designs make 4 thumbnails and use 4 credits.

A Prompt run has no title, so the thumbnail's own page takes its name from the first line of the description, and a first line that says what the video is about doubles as a good name.

Once a picture is finished it can be changed in words or by marking the parts to change, and a 16:9 picture can be given a Shorts version.

Prompt or Recreate

When to describe it, and when to show it.

Prompt suits a picture you can already see and say: a person, a place, an object or two and a short headline, above all a scene that could have been photographed.

When the look you want is easier to point at than to put into words, such as the layout, colors and lettering of a thumbnail on another channel, start in Recreate instead. Paste that video's link or upload the image, and the thumbnail becomes the reference a new one is built on.

Both ways in lead to the same Persona tile, the same frames and the same Make button, and nothing is used until you press it, so trying one and then the other costs nothing.

Describe your next thumbnail.

Open the Studio in Prompt, pick Demo or a persona of your own, and write what the picture should show.

Open the Studio

2 free credits, no card