Thumbnail Layout Brief Prompt for Designers

Start here

Paste your video topic (one or two sentences) and the exact candidate title into an ordinary AI chat that accepts text. Add anything optional you already have: audience, platform, aspect ratio, forbidden elements, brand colours, or a plain description of a photo or person the thumbnail may use.

Then paste the complete prompt appended below. The model should first restate your topic and title, then list missing inputs, then give three layouts and a comparison line. If the topic or title is missing, it should stop and ask for both.

Check the output by confirming each of these:

• Every title word in every layout matches your words, in order, with nothing added or cut.
• The three directions differ in composition and focal point, not only colour.
• Each text block sits inside the stated safe area.
• No statistic, colour, number or brand appears that you did not supply, unless it is clearly labelled an assumption.

What this prompt does

It produces a written creative brief, not an image. Each of the three layouts names a direction, describes canvas and safe area, focal element with an approximate grid position, background layers, exact text placement, a hierarchy note, and two risks with quick fixes. A short comparison line then says which layout suits which mood or platform.

Fictional example of the output shape (not a real test): for topic "a beginner guide to making cold-brew coffee at home" and title "Cold Brew, No Fuss," one direction might centre a glass jug with the title as two short lines bottom-left, while another splits the frame into a photo half and a flat colour block holding the full title.

Inputs and permissions

What you must supply:

If no platform or aspect ratio is given, the prompt should offer two common options and label which layout fits each instead of assuming one. The prompt is a text-only instruction for a person to execute; it should not claim it created an image, opened an editor, exported a file or posted anything, and it should not claim a design was tested with viewers.

Privacy and limitations

Anything you paste goes to whichever chat service you use, so avoid pasting confidential brand assets, unreleased product details or personal data you would not share there. The output is a starting brief for a human designer, not a finished thumbnail, and it will not verify legibility in practice.

This guide describes documented setup for the prompt. Source reviewed; runtime not tested.

FAQ

Do I need any software or account setup? No. Paste the prompt into an ordinary AI chat that accepts text. No terminal or API setup is involved.

What if I only have a rough title? The prompt keeps the exact words you supply and does not rewrite them, so decide the final wording first. If the prompt splits a long title across lines, it must keep every word in order.

Complete reusable prompt

You are helping me brief a graphic designer for a video thumbnail. I will give you a subject and a candidate title. Work only from what I supply. Do not invent product names, logos, statistics, faces, brand colours or claims that I have not stated.

INPUT I WILL PROVIDE

• Video topic: one or two sentences describing what the video covers.
• Candidate title: the exact title text, or the words that must appear in the image.
• Optional: audience, platform, aspect ratio, forbidden elements, brand colours.
• Optional: a plain description of any photo or person the thumbnail may use.

YOUR TASK Produce three clearly different thumbnail layout briefs for the SAME topic and title, so the designer can choose a direction. Vary composition, focal point and how the text sits, not just the colour mood. Each brief must be a complete instruction a designer could follow without seeing the video.

OUTPUT FORMAT For each of the three layouts, use this structure:

1. Direction name and one-line intent.
2. Canvas and safe area: state the aspect ratio you were given, or ask for it; describe where the crop-safe zone sits and keep all text inside it.
3. Focal element: what the eye lands on first, described in words (subject, object, scene), with its approximate position using a simple grid or fractions.
4. Background and depth: describe layers from back to front. No invented photos after I have named none.
5. Text placement: exact title words, the block they occupy, alignment, and whether the title is the main or secondary element. If the title is long, say how to split it across lines and keep the wording unchanged.
6. Hierarchy note: one sentence on what the viewer reads first, second and third.
7. Two risks: one legibility risk and one credibility risk for this direction, each with a quick fix. Then add a short comparison line saying which layout suits which mood or platform, based only on the directions you wrote.

RULES

• Do not rewrite, shorten or embellish my title. Keep the exact words.
• Do not add numbers, awards, prices or comparisons unless I supplied them.
• Do not name any software, template pack or platform feature.
• If I have not given a platform or aspect ratio, offer two common options and label which layout fits each, rather than assuming one.
• If my title has more than six words, propose the line breaks you would use and keep every word.
• Keep the whole answer readable, not padded.

UNCERTAINTY AND MISSING INPUT Before the three layouts, list any missing inputs that would change your placement or hierarchy. If the topic or title is absent, stop and ask for both rather than guessing. If you are unsure whether a described element suits the audience, mark it as an assumption.

CHECKS BEFORE YOU FINISH

• Every title word in each layout matches the words I gave, in order.
• Each of the three directions differs in at least composition and focal point.
• Every text block sits inside the stated safe area.
• No element, statistic or colour appears that I did not supply or that you did not clearly label as an assumption.

BOUNDARIES This is a written creative brief only. Do not claim you created an image, opened an editor, exported a file or posted anything. Do not claim any design has been tested with viewers. You are preparing instructions for a person to execute.

WORKED MINI EXAMPLE (fictional, for shape only) Input: Topic: a beginner guide to making cold-brew coffee at home. Title: Cold Brew, No Fuss. Platform: 16:9 video site. Illustrative output shape: Direction 1, "Object Hero," centred glass jug on a plain surface, title as two short lines bottom-left inside the safe area, clean and calm. Direction 2, "Split Frame," left half showing the brewing jar, right half a flat colour block holding the full title, bolder and more direct. Direction 3, "Before and After," dry grounds on one side and finished drink on the other, title across the lower third with the two key words emphasised by size only, wording unchanged. Pass and fail checks for this example: Pass if all three keep "Cold Brew, No Fuss" exactly and place it inside the safe area, and if the three differ in composition. Fail if a direction invents a coffee brand, a health claim, or a number of minutes not supplied, or if the title wording is changed.

Start by restating the topic and title you received, then list missing inputs, then give the three layouts and the comparison line.

References and reuse

• ChatGPT release notes

Original TokRepo prompt · CC BY 4.0. Reference documents retain their own rights.