Skip to content
PromptifyLab

Creators & Social

Thumbnails & Social Visuals prompts

7 prompts Free · no sign-up Works in ChatGPT, Claude & Gemini Search & filter these

Thumbnail concepts, consistent brand styles, graphics, banners and alt text. Below are 7 copy-ready prompts. Fill in the [BRACKETS], copy, and paste into ChatGPT, Claude, Gemini or any capable assistant.

A thumbnail or social graphic has less than a second to work. The prompts here start from the idea – the one emotion or curiosity the image must create – then produce layouts and image prompts, a consistent style block for your whole feed, and a critique before you publish.

The 7 prompts

Beginner 4 blanks to fill

Design YouTube thumbnail concepts that match the title

Generate thumbnail concepts and an image prompt for each.

Prompt
Create YouTube thumbnail concepts.

VIDEO TITLE: [TITLE]
WHAT HAPPENS IN THE VIDEO / THE PAYOFF: [PAYOFF]
MY CHANNEL STYLE: [colours, face or faceless, usual layout]
IMAGE TOOL: [e.g. ChatGPT images, Gemini, Midjourney, Firefly, Photoshop]

Give me 4 concepts:
For each:
1. THE IDEA in one sentence (what emotion or curiosity it triggers)
2. LAYOUT - main subject position, background, one focal point
3. TEXT - 0 to 4 words max, that ADD to the title rather than repeat it
4. COLOURS - high contrast, and different from what usually appears around it
5. IMAGE PROMPT - for the background/subject, 16:9, with space left for text
6. SMALL-SIZE TEST - describe what a viewer sees at phone thumbnail size

Rules:
- One focal point. Three things = nothing.
- No misleading images (showing something that is not in the video).
- No real people's likeness unless it is me, and no brand logos I do not own.

What you get: Four thumbnail concepts with layout, text, colours, an image prompt and a small-size readability check.

Tip: Test at small size. Shrink your thumbnail to the size of a fingernail - if you cannot tell what it is, simplify.

Open in Written for ChatGPT, Claude, Gemini · Reviewed September 25, 2026
Intermediate 4 blanks to fill

Create a consistent visual style for your channel or brand

Build a reusable style guide and prompt block for all your social visuals.

Prompt
Create a visual style guide for my social media.

BRAND / CHANNEL: [DESCRIBE]
AUDIENCE: [WHO]
FEELING I WANT: [e.g. trustworthy, playful, luxury, energetic]
PLATFORMS: [LIST]
EXAMPLES I LIKE (describe, do not name living artists): [DESCRIBE]

Produce:
1. COLOUR PALETTE - 5 colours with hex codes and when to use each
2. TYPE - 2 font pairings (free fonts) and sizes for thumbnails vs carousels
3. IMAGE STYLE - photography or illustration rules (lighting, angle, background)
4. LAYOUT TEMPLATES - 3 reusable layouts per platform
5. A STYLE PROMPT BLOCK - a paragraph I can add to any AI image prompt to keep images consistent
6. DO / DON'T list with 5 each

What you get: A palette, font pairings, image rules, layout templates, a reusable style prompt block and do/don't lists.

Tip: The style prompt block is the practical output. Paste it into every image prompt and your feed starts looking like one brand.

Open in Written for ChatGPT, Claude, Gemini · Reviewed September 25, 2026
Beginner 3 blanks to fill

Write image prompts for quote and tip graphics

Generate backgrounds and layouts for text-based social posts.

Prompt
Create quote or tip graphics for social media.

TEXT FOR THE GRAPHICS (one per graphic): [LIST QUOTES OR TIPS - only quotes I can attribute correctly]
PLATFORM AND SIZE: [e.g. Instagram 4:5, Pinterest 2:3, LinkedIn 1:1]
BRAND STYLE: [COLOURS, FONTS, MOOD]

For each graphic:
- Background image prompt (abstract, texture, scene) with clear space for text
- Where the text goes and suggested font size ratio
- Colour of text for contrast
- Caption to post with it

Rules: check every quote is attributed to the right person - misattributed quotes are very common. If I did not give a source for a quote, mark it [VERIFY ATTRIBUTION].

What you get: Per-graphic background prompts, text placement, contrast colours, captions and quote-attribution checks.

Tip: Misattributed quotes are everywhere online. Verify before posting - it is an easy credibility loss.

Open in Written for ChatGPT, Claude, Gemini · Reviewed September 25, 2026
Beginner 2 blanks to fill

Critique a thumbnail or social graphic

Get specific, blunt feedback on a thumbnail or graphic before publishing.

Prompt
Critique my thumbnail / social graphic. (Attach or describe the image.)

WHAT IT IS FOR: [PLATFORM, POST OR VIDEO TITLE]
WHAT I WANT A VIEWER TO FEEL OR DO: [GOAL]

Evaluate:
1. Focal point - is there one clear thing to look at?
2. Readability at small size
3. Contrast and colour
4. Text - number of words, size, whether it repeats the title
5. Faces/emotion (if any) - clear expression?
6. Clutter - list anything that could be removed
7. Honesty - does it match the content?

Give a score out of 10, the top 3 fixes, and a description of an improved version.

What you get: A seven-point critique with a score, top three fixes and a description of an improved version.

Tip: Attach the actual image if your AI tool supports images; descriptions lose the details that matter.

Open in Written for ChatGPT, Claude, Gemini · Reviewed September 25, 2026
Intermediate 4 blanks to fill

Make a set of matching images for a carousel or series

Keep characters, colours and style consistent across multiple AI images.

Prompt
Help me create a consistent set of AI images.

WHAT THE SET IS FOR: [e.g. 8-slide carousel, story series, product line]
SUBJECT / CHARACTER: [DESCRIBE IN DETAIL - age range, clothing, colours, distinctive features]
STYLE: [DESCRIBE]
IMAGE TOOL: [TOOL]

Produce:
1. A MASTER DESCRIPTION of the subject to paste into every prompt, word for word
2. A STYLE BLOCK to paste into every prompt
3. PROMPTS for each image in the set, changing only the action, setting or angle
4. TOOL TIPS - if my tool supports reference images, character references or seeds, how to use them for consistency
5. A CHECKLIST to compare the images for drift (clothing, colours, face, proportions)

What you get: A master subject description, style block, per-image prompts, consistency tool tips and a drift checklist.

Tip: Change one thing per image. If the action, setting and angle all change at once, the character drifts too.

Open in Written for ChatGPT, Claude, Gemini · Reviewed September 25, 2026
Beginner 4 blanks to fill

Create a profile picture and banner set

Design profile images and banners that work across platforms.

Prompt
Design my profile picture and banner set.

WHO / WHAT: [PERSON OR BRAND]
PLATFORMS: [YouTube, Facebook, X, LinkedIn, TikTok...]
STYLE: [photo / illustrated / logo-based]
MESSAGE FOR THE BANNER: [WHAT YOU DO, UPLOAD SCHEDULE, TAGLINE]

Give me:
1. Profile picture concept that reads at tiny size (circle crop safe)
2. Banner concept per platform, noting the area that is visible on mobile vs desktop and keeping key text in the safe centre
3. Image prompts for backgrounds
4. Text for each banner (short)
5. A checklist to preview on phone before saving

Tell me to check each platform's current recommended dimensions, as they change.

What you get: A profile concept, per-platform banner concepts with safe areas, background prompts and banner text.

Tip: Banners are cropped differently on phones, TVs and desktops. Keep all text in the middle strip.

Open in Written for ChatGPT, Claude, Gemini · Reviewed September 25, 2026
Beginner 3 blanks to fill

Write alt text and captions for social images

Make social images accessible with useful alt text.

Prompt
Write alt text for my social media images.

IMAGES (describe each or attach): [LIST]
PLATFORM: [PLATFORM]
CONTEXT OF THE POST: [WHAT THE POST IS ABOUT]

For each image write alt text that:
- Describes what matters for understanding the post, in under 125 characters where possible
- Includes any text in the image
- Does not start with "Image of" or "Picture of"
- Does not stuff keywords

Also note which platforms let me add alt text and where the setting is (tell me to check, as menus change).

What you get: Concise, useful alt text for each image, including any embedded text.

Tip: If the image contains text, that text must be in the alt text. Screen readers cannot read words inside images.

Open in Written for ChatGPT, Claude, Gemini · Reviewed September 25, 2026

Where AI actually helps here

  • Generating several thumbnail concepts that complement the title
  • Writing a reusable style block so every AI image matches
  • Critiquing a graphic for clutter and small-size readability
  • Writing alt text

Where it falls down

  • Rendering exact text – add words in an editor
  • Guaranteeing consistent faces across images without reference tools
  • Knowing whether an image resembles a trademark

The mistake almost everyone makes: Three focal points

When a thumbnail has a face, an object, a logo and five words, viewers see nothing. Ask for one focal point and test the design at fingernail size before you publish.

Free tool: Image Prompt Builder

Runs in your browser. No sign-up, nothing uploaded.

Open the Image Prompt Builder →

Questions people ask


Can I make YouTube thumbnails with AI?

Yes – for backgrounds, scenes and concepts. Add your own photo and text in an editor for accuracy and consistency.


What size is a YouTube thumbnail?

YouTube recommends a 16:9 image; check YouTube Help for current dimensions and file-size limits.


How do I keep AI images consistent?

Reuse the same subject description and style block in every prompt, and use reference image features where your tool has them.