Claude Prompt Library

30 Claude Prompts That Design Thumbnails

30 copy-paste prompts

Describe the video, episode, or article and Claude returns a finished thumbnail as a self-contained HTML/CSS or SVG artifact you can preview and screenshot instantly. Prompts for YouTube, podcast covers, blog and course art, A/B variants, and bold text-overlay layouts. Not "give me some ideas."

In short: This page contains 30 copy-paste ready prompts, organized into 6 categories with a description and pro tip for each. The first 15 prompts are free instantly โ€” no signup needed. Hand-curated and tested by the AI Academy team.

By Louis Corneloup ยท Founder, Techpresso
Last updated ยทHand-curated & tested by the AI Academy team

YouTube Thumbnails

5 prompts

High-CTR Talking-Head Thumbnail

1/30

You are a YouTube thumbnail designer who studies CTR data for a living. <context> I need a 1280x720 talking-head video thumbnail built as one self-contained HTML file with inline CSS, scaled to a 16:9 frame so I can preview it instantly as an artifact and screenshot it. </context> <inputs> - Video topic / hook: [WHAT THE VIDEO PROMISES] - The 3-5 word overlay text: [BIG PUNCHY PHRASE] - Channel niche and audience: [E.G. PERSONAL FINANCE, BEGINNERS] - Emotion the face should convey: [SHOCKED / EXCITED / SERIOUS] - Brand colors: [HEX OR DESCRIBE] </inputs> <task> Build a 1280x720 thumbnail at proper aspect ratio: a face placeholder block on one side (labeled with the target expression), a bold high-contrast text overlay of the 3-5 word phrase with a heavy outline or background plate so it stays readable at small sizes, a supporting visual element or icon, and a subtle directional cue (arrow, circle, or glow) pulling the eye to the subject. Use a strong color contrast between text and background. </task> <constraints> - One self-contained HTML file, exactly 16:9, no external assets except Google Fonts. - Text must stay legible at 320px wide (mobile feed); max two text colors. - No lorem ipsum; use the real overlay phrase. Accessible contrast on the text plate. </constraints> <format> Return the full HTML in an artifact, then explain the contrast and focal-point choices and which element earns the click. </format>

Builds a 1280x720 talking-head YouTube thumbnail with a bold readable text overlay and focal cue as a previewable artifact.

๐Ÿ’ก

Pro tip: Tell Claude the single most-clickable word in your overlay and ask it to make that one word the largest and brightest element.

Listicle / Number Thumbnail

2/30

You are a YouTube growth designer specializing in list-video thumbnails. <context> I need a 1280x720 thumbnail for a list video, built as one self-contained HTML file with inline CSS at 16:9 so I can preview it as an artifact and export it. </context> <inputs> - Big number to feature: [E.G. 7, 12, 100] - What the number counts: [E.G. AI TOOLS, MISTAKES, HACKS] - Short benefit phrase: [WHY WATCH] - Niche and vibe: [DESCRIBE] - Color scheme: [HEX OR DESCRIBE] </inputs> <task> Build a thumbnail where the big number is the dominant hero element (oversized, styled with a gradient or outline), paired with a short two-to-four word label of what it counts and a small benefit line. Add a simple supporting graphic or icon block. Use a layout that reads instantly: number first, label second. </task> <constraints> - One self-contained HTML file, exactly 16:9, Google Fonts only. - The number must dominate visually and stay crisp at small sizes; no more than three type sizes. - High contrast, accessible text, no filler copy. </constraints> <format> Return the full HTML in an artifact, then explain why you sized the number the way you did and one variant to test. </format>

Generates a list-video YouTube thumbnail with an oversized hero number and benefit label as a previewable artifact.

๐Ÿ’ก

Pro tip: Ask Claude to make the number a different color from everything else so it pops as the single anchor of the design.

Before / After Split Thumbnail

3/30

You are a YouTube thumbnail designer who builds high-contrast transformation visuals. <context> I need a 1280x720 before-and-after split thumbnail built as one self-contained HTML file with inline CSS at 16:9, previewable instantly as an artifact. </context> <inputs> - The transformation: [E.G. MESSY ROOM TO CLEAN, BROKE TO RICH, OLD UI TO NEW] - Left (before) label: [SHORT WORD] - Right (after) label: [SHORT WORD] - Overlay headline: [3-5 WORDS] - Color tone for each side: [DESCRIBE] </inputs> <task> Build a split-screen thumbnail divided by a clear diagonal or vertical line: a desaturated or dull left half labeled "before" and a vibrant right half labeled "after," each with a placeholder visual block. Overlay the short headline across the seam with a heavy text treatment, plus a small arrow or "VS" marker on the divide. </task> <constraints> - One self-contained HTML file, exactly 16:9, Google Fonts only. - The two halves must feel clearly distinct (color, brightness); divider is crisp. - Headline legible over both halves at small sizes; accessible contrast. </constraints> <format> Return the full HTML in an artifact, then explain how you signaled the contrast and where the eye lands first. </format>

Builds a before-and-after split-screen YouTube thumbnail with a clear divider and headline as a previewable artifact.

๐Ÿ’ก

Pro tip: Tell Claude to make the after side brighter and more saturated than the before side so the payoff reads in a half-second.

Tutorial / How-To Thumbnail

4/30

You are a thumbnail designer who specializes in educational and tutorial channels. <context> I need a 1280x720 tutorial thumbnail built as one self-contained HTML file with inline CSS at 16:9, previewable instantly as an artifact. </context> <inputs> - What the tutorial teaches: [SKILL OR OUTCOME] - Tool / app shown: [NAME, IF ANY] - Overlay text: [3-6 WORDS, E.G. "DO THIS IN 5 MIN"] - Audience level: [BEGINNER / PRO] - Brand colors: [HEX OR DESCRIBE] </inputs> <task> Build a clean, trustworthy tutorial thumbnail: a UI or screen placeholder block on one side (labeled with the tool), a clear instructional overlay phrase with a step or arrow cue, and a small badge like "step by step" or a time stamp. Keep it less hypey than entertainment thumbnails but still high-contrast and scannable. </task> <constraints> - One self-contained HTML file, exactly 16:9, Google Fonts only. - Layout reads as helpful and clear, not clickbait; text legible at small sizes. - Accessible contrast; no filler copy. </constraints> <format> Return the full HTML in an artifact, then explain how the design signals credibility and one tweak for higher CTR. </format>

Creates a clear, trustworthy how-to YouTube thumbnail with a tool screen and instructional overlay as a previewable artifact.

๐Ÿ’ก

Pro tip: Add a small time badge like "5 min" โ€” Claude will place it as a corner sticker that promises a quick, low-effort win.

Reaction / Drama Thumbnail

5/30

You are a thumbnail designer for commentary, reaction, and drama videos. <context> I need a 1280x720 high-emotion reaction thumbnail built as one self-contained HTML file with inline CSS at 16:9, previewable instantly as an artifact. </context> <inputs> - The topic / drama: [WHAT HAPPENED] - Reaction face expression: [SHOCKED / DISGUSTED / LAUGHING] - Overlay text: [SHORT REACTIVE PHRASE] - Secondary subject or screenshot: [WHAT IT REACTS TO] - Color mood: [DESCRIBE] </inputs> <task> Build a reaction thumbnail with a large expressive face placeholder on one side (labeled with the expression), a contrasting subject or screenshot block on the other, an emoji or symbol overlay (red circle, arrow, exclamation) drawing tension, and a punchy reactive overlay phrase. Use saturated, urgent color and a slight tilt or glow for energy. </task> <constraints> - One self-contained HTML file, exactly 16:9, Google Fonts only. - High visual tension but text stays readable at small sizes; max two text colors. - Accessible contrast on the headline; no filler. </constraints> <format> Return the full HTML in an artifact, then explain the tension devices you used and how to keep it on-brand. </format>

Generates a high-emotion reaction or drama YouTube thumbnail with expressive face and tension cues as a previewable artifact.

๐Ÿ’ก

Pro tip: Give Claude the exact emotion in plain words; the labeled expression on the face block drives the whole composition.

XML tags are just the start. Learn the full Claude workflow.

A growing library of 300+ hands-on AI tutorials covering Claude, ChatGPT, and 50+ tools. New tutorials added every week.

Start 7-Day Free Trial

Podcast & Audio Cover Art

5 prompts

Square Podcast Cover Art

6/30

You are a podcast brand designer who builds covers that pop in directory grids. <context> I need a 3000x3000-ratio podcast cover built as one self-contained HTML file with inline CSS scaled to a 1:1 frame, previewable instantly as an artifact and exportable as a square image. </context> <inputs> - Show name: [EXACT TITLE] - Tagline (optional): [SHORT LINE] - Topic / genre: [E.G. STARTUPS, TRUE CRIME, COMEDY] - Mood and color palette: [DESCRIBE OR HEX] - Host vibe: [SOLO, INTERVIEW, PANEL] </inputs> <task> Build a square cover with the show name as the dominant element, set in a confident type treatment, plus an optional tagline and a simple genre-fitting graphic motif (waveform, mic, abstract shape) using CSS or inline SVG. Keep it bold and legible even as a tiny tile in Apple Podcasts and Spotify. Use a cohesive two-to-three color palette. </task> <constraints> - One self-contained HTML file, exactly 1:1, Google Fonts and inline SVG only. - Title must be readable at 150x150px (directory thumbnail size); no thin fonts. - Accessible contrast; no filler text. </constraints> <format> Return the full HTML in an artifact, then explain the palette and type choices and how it holds up at tiny sizes. </format>

Builds a bold square podcast cover that stays legible at directory-tile size as a previewable artifact.

๐Ÿ’ก

Pro tip: Ask Claude to preview the title at 150px and shrink any element that disappears โ€” directory tiles are brutally small.

Episode-Specific Cover Variant

7/30

You are a podcast designer who creates per-episode cover variants on a fixed brand template. <context> I have a podcast and need an episode-specific square cover built as one self-contained HTML file with inline CSS at 1:1, previewable as an artifact, that keeps my brand frame but swaps the episode content. </context> <inputs> - Show name and brand colors: [TITLE PLUS HEX] - Episode number and title: [E.G. EP 42 โ€” SCALING TO 1M USERS] - Guest name (optional): [NAME AND TITLE] - Episode theme keyword: [ONE WORD OR PHRASE] - Reusable brand motif: [WHAT STAYS CONSTANT] </inputs> <task> Build an episode cover that keeps a consistent brand band (logo lockup or name strip) and swaps in: the episode number badge, the episode title, an optional guest photo placeholder with name, and a theme accent color. Design it so the template is obviously reusable across episodes by only changing the variable zone. </task> <constraints> - One self-contained HTML file, exactly 1:1, Google Fonts and inline SVG only. - Clearly separate the fixed brand zone from the per-episode variable zone in the markup with comments. - Legible at small sizes; accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then label which CSS variables to change for the next episode. </format>

Creates a reusable per-episode podcast cover template with a fixed brand zone and swappable content as a previewable artifact.

๐Ÿ’ก

Pro tip: Have Claude expose the episode title, number, and accent color as CSS variables so producing the next cover is a 30-second edit.

Audiogram / Quote Clip Card

8/30

You are a social audio designer who builds shareable audiogram cards for podcast clips. <context> I need a square audiogram-style quote card built as one self-contained HTML file with inline CSS at 1:1, previewable instantly as an artifact, made to be screenshotted or used behind an audio waveform. </context> <inputs> - Pull quote from the episode: [THE QUOTE] - Speaker name and title: [WHO SAID IT] - Show name: [TITLE] - Brand colors: [HEX OR DESCRIBE] - Vibe: [BOLD, MINIMAL, WARM] </inputs> <task> Build a quote card: the pull quote set large and well-broken across lines for readability, a speaker attribution line, a small show lockup, and a static waveform graphic (CSS bars or inline SVG) along the bottom suggesting the audio. Keep generous margins and strong typographic hierarchy so the quote is the hero. </task> <constraints> - One self-contained HTML file, exactly 1:1, Google Fonts and inline SVG only. - Quote text wraps cleanly with no orphaned single words; readable on social feeds. - Accessible contrast; no filler beyond the real quote. </constraints> <format> Return the full HTML in an artifact, then explain the line-break and hierarchy decisions for the quote. </format>

Generates a shareable audiogram quote card with a hero pull quote and static waveform as a previewable artifact.

๐Ÿ’ก

Pro tip: Paste the exact quote and let Claude control the line breaks โ€” good breaks are what make a quote card feel designed, not dumped.

Interview Cover With Guest Photo

9/30

You are a podcast cover designer who features high-profile interview guests. <context> I need a square interview-episode cover built as one self-contained HTML file with inline CSS at 1:1, previewable as an artifact, that puts the guest front and center. </context> <inputs> - Show name: [TITLE] - Guest name and credential: [NAME, WHY THEY MATTER] - Conversation hook: [WHAT THIS EPISODE PROMISES] - Brand color and accent: [HEX] - Layout preference: [GUEST LEFT / RIGHT / CENTER] </inputs> <task> Build a cover with a prominent guest photo placeholder (labeled with the guest name), the guest's name in large type, a one-line credential, a short conversation hook, and the show lockup. Use a layout that makes the guest the visual draw while keeping the show recognizable. </task> <constraints> - One self-contained HTML file, exactly 1:1, Google Fonts and inline SVG only. - Guest name and show name both legible at small sizes; clear photo placeholder caption. - Accessible contrast; no filler copy. </constraints> <format> Return the full HTML in an artifact, then explain how you balanced guest prominence against show branding. </format>

Builds an interview-episode podcast cover that features the guest while keeping the show recognizable as a previewable artifact.

๐Ÿ’ก

Pro tip: Tell Claude how famous the guest is; a marquee name should dwarf the show name, an unknown guest should not.

YouTube-Video Podcast Thumbnail

10/30

You are a designer who adapts podcast branding into 16:9 YouTube thumbnails. <context> My podcast also publishes on YouTube. I need a 1280x720 video thumbnail built as one self-contained HTML file with inline CSS at 16:9, previewable as an artifact, that translates my square cover branding into a wide format. </context> <inputs> - Show name and brand colors: [TITLE PLUS HEX] - Episode title: [EPISODE HEADLINE] - Host and guest faces: [WHO APPEARS] - The episode hook: [WHY CLICK] - Vibe: [DESCRIBE] </inputs> <task> Build a 16:9 thumbnail with two face placeholders (host and guest, labeled), a bold episode headline overlay, the show lockup as a consistent brand band, and an accent treatment that ties back to the square cover. Make it feel like the same brand but optimized for the wide YouTube feed. </task> <constraints> - One self-contained HTML file, exactly 16:9, Google Fonts and inline SVG only. - Faces and headline legible at small sizes; show band consistent and recognizable. - Accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then explain how you carried the square-cover branding into 16:9. </format>

Adapts a podcast's square branding into a 16:9 YouTube video thumbnail with host and guest as a previewable artifact.

๐Ÿ’ก

Pro tip: Reuse the same accent color and font from your square cover so subscribers recognize the show across both platforms.

Blog, Article & Content Thumbnails

5 prompts

Blog Post Featured Image

11/30

You are an editorial designer who creates featured images for blog articles. <context> I need a 1200x630 blog featured image built as one self-contained HTML file with inline CSS at that aspect ratio, previewable instantly as an artifact and exportable for the post header and social share. </context> <inputs> - Article title: [EXACT HEADLINE] - Topic / category: [E.G. PRODUCTIVITY, MARKETING] - Blog brand colors and fonts: [DESCRIBE OR HEX] - Tone: [PROFESSIONAL, FRIENDLY, TECHNICAL] - Optional author or logo: [NAME OR BRAND] </inputs> <task> Build a clean featured image: the article title set in a strong editorial type treatment as the hero, a category tag or eyebrow label, a subtle background pattern or shape motif, and an optional small brand or author lockup. Keep it readable as both a 1200px header and a smaller social preview. </task> <constraints> - One self-contained HTML file, 1200x630 ratio, Google Fonts and CSS/SVG shapes only. - Title legible at social-preview size; no busy backgrounds competing with text. - Accessible contrast; use the real headline, no placeholder. </constraints> <format> Return the full HTML in an artifact, then explain the type hierarchy and how to reuse it as a template for other posts. </format>

Builds a 1200x630 blog featured image with an editorial title treatment and brand motif as a previewable artifact.

๐Ÿ’ก

Pro tip: Give Claude your blog's two brand fonts so every featured image looks like part of the same publication.

Open Graph Social-Share Card

12/30

You are a designer who builds Open Graph cards for link previews on social platforms. <context> I need a 1200x630 Open Graph share card built as one self-contained HTML file with inline CSS at that ratio, previewable as an artifact, optimized for how it appears when a link is pasted into X, LinkedIn, and Slack. </context> <inputs> - Page / article title: [EXACT TITLE] - Site or brand name: [NAME] - One-line description or hook: [SHORT LINE] - Brand colors and logo: [HEX, LOGO TEXT] - Content type: [ARTICLE / PRODUCT / TOOL] </inputs> <task> Build an OG card with the title as the dominant element, a brief supporting line, a consistent brand lockup (logo text plus accent bar), and a clean background that survives the heavy crop and compression of social previews. Prioritize a layout that still reads when scaled down in a feed. </task> <constraints> - One self-contained HTML file, exactly 1200x630, Google Fonts and CSS/SVG only. - Keep critical content out of the extreme edges (safe margins for cropping). - High contrast, accessible, no filler copy. </constraints> <format> Return the full HTML in an artifact, then explain the safe-area and contrast decisions for social previews. </format>

Generates a 1200x630 Open Graph link-preview card with safe margins and brand lockup as a previewable artifact.

๐Ÿ’ก

Pro tip: Ask Claude to keep all text inside a centered safe zone โ€” feeds crop the edges differently on every platform.

Pinterest Tall Pin Graphic

13/30

You are a Pinterest designer who builds vertical pins that get saved. <context> I need a 1000x1500 vertical Pinterest pin built as one self-contained HTML file with inline CSS at 2:3, previewable instantly as an artifact. </context> <inputs> - Pin topic / promise: [WHAT IT OFFERS] - Big overlay headline: [SAVE-WORTHY PHRASE] - Brand name or URL: [TO PLACE AT BOTTOM] - Niche aesthetic: [E.G. COZY, MINIMAL, BOLD] - Color palette: [HEX OR DESCRIBE] </inputs> <task> Build a tall pin: a strong top-to-bottom layout with a large benefit headline in the upper or center band, a supporting image placeholder zone, a value or list cue (e.g. three quick bullets or a numbered teaser), and a brand or URL footer. Design for the vertical feed where pins are scrolled fast. </task> <constraints> - One self-contained HTML file, exactly 2:3 vertical, Google Fonts and CSS/SVG only. - Headline dominant and readable on mobile; brand URL legible but not loud. - Accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then explain why the vertical hierarchy earns the save. </format>

Builds a 1000x1500 vertical Pinterest pin with a save-worthy headline and brand footer as a previewable artifact.

๐Ÿ’ก

Pro tip: Put a numbered or list cue near the top; Pinterest savers love pins that promise a quick, scannable payoff.

Newsletter Header Banner

14/30

You are a designer who creates email newsletter header banners. <context> I need a wide newsletter header banner built as one self-contained HTML file with inline CSS at roughly 1200x300, previewable as an artifact, to sit at the top of an email issue. </context> <inputs> - Newsletter name: [TITLE] - Issue number or date: [E.G. ISSUE #18] - Theme of this issue: [TOPIC] - Brand colors and font feel: [DESCRIBE OR HEX] - Personality: [SHARP, FUN, PREMIUM] </inputs> <task> Build a header banner with the newsletter name as a confident wordmark, an issue number or date line, an optional one-line theme, and a brand motif (shape, line, or pattern) that ties issues together. Keep it short and wide, readable on both desktop and mobile email clients. </task> <constraints> - One self-contained HTML file, ~1200x300 ratio, Google Fonts and CSS/SVG only. - Wordmark legible on a phone; no thin fonts that break in email rendering. - Accessible contrast; reusable across issues with a swappable issue number. </constraints> <format> Return the full HTML in an artifact, then label the swappable parts so I can reuse it each issue. </format>

Creates a reusable wide newsletter header banner with a wordmark and issue line as a previewable artifact.

๐Ÿ’ก

Pro tip: Ask Claude to make the issue number a separate swappable element so you reuse the same header for every send.

Blog Category / Series Banner

15/30

You are a content designer who builds consistent banners for a blog content series. <context> I need a set-up-as-template series banner built as one self-contained HTML file with inline CSS at 1200x630, previewable as an artifact, that looks like part of an ongoing series. </context> <inputs> - Series name: [E.G. "BUILD IN PUBLIC"] - Part number and topic: [PART 3 โ€” PRICING] - Brand colors and font: [DESCRIBE OR HEX] - Series motif idea: [SHAPE, BADGE, OR PATTERN] - Tone: [DESCRIBE] </inputs> <task> Build a series banner with a fixed series badge or label, a part-number indicator, the specific entry's topic as the headline, and a recurring visual motif that signals it belongs to the series. Make the fixed series elements visually distinct from the per-entry variable elements. </task> <constraints> - One self-contained HTML file, exactly 1200x630, Google Fonts and CSS/SVG only. - Series identity stays constant across entries; only the part number and topic change. - Accessible contrast; no filler; comment the variable zone. </constraints> <format> Return the full HTML in an artifact, then explain which elements stay fixed and which change per entry. </format>

Builds a templated blog series banner with a fixed series identity and swappable entry topic as a previewable artifact.

๐Ÿ’ก

Pro tip: Lock the series badge and motif as constants; only let the part number and topic vary so the whole series feels like one set.

Course & Product Thumbnails

5 prompts

Online Course Card Thumbnail

16/30

You are a course-platform designer who builds thumbnails for online lessons and courses. <context> I need a 16:9 course thumbnail built as one self-contained HTML file with inline CSS, previewable as an artifact, styled to look polished in a course catalog grid. </context> <inputs> - Course title: [EXACT TITLE] - What students learn / outcome: [SKILL OR RESULT] - Level: [BEGINNER / INTERMEDIATE / ADVANCED] - Brand colors and font: [DESCRIBE OR HEX] - Optional instructor or logo: [NAME OR BRAND] </inputs> <task> Build a course thumbnail with the title as the hero, a level badge, a short outcome line, a topic-fitting icon or graphic motif, and an optional instructor or brand lockup. Design it to look credible and premium in a catalog grid alongside other courses. </task> <constraints> - One self-contained HTML file, exactly 16:9, Google Fonts and CSS/SVG only. - Title legible at catalog-card size; consistent style for a full course library. - Accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then explain how to keep a whole course library visually consistent. </format>

Builds a polished 16:9 online course thumbnail with title, level badge, and outcome line as a previewable artifact.

๐Ÿ’ก

Pro tip: Tell Claude the level and outcome up front; a clear "beginner" badge plus a concrete result is what makes a card click-worthy.

Lesson / Module Thumbnail Set

17/30

You are a designer who creates a consistent thumbnail system for course modules. <context> I need a single 16:9 lesson thumbnail built as one self-contained HTML file with inline CSS, previewable as an artifact, designed as a reusable system so every module in the course matches. </context> <inputs> - Course name and brand colors: [TITLE PLUS HEX] - Module number and title: [MODULE 4 โ€” THE PRICING MODEL] - Section or chapter label: [OPTIONAL] - Icon idea for this module: [WHAT REPRESENTS IT] - Visual style: [DESCRIBE] </inputs> <task> Build a module thumbnail with a consistent course frame (course name strip plus accent), a prominent module number, the module title, an optional section label, and a swappable module icon. Make it obviously part of a numbered system so the course library reads as a cohesive set. </task> <constraints> - One self-contained HTML file, exactly 16:9, Google Fonts and CSS/SVG only. - Fixed frame stays constant; module number, title, and icon are the variable zone (commented). - Accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then list the CSS variables to change for each subsequent module. </format>

Creates a reusable course-module thumbnail system with a fixed frame and swappable module content as a previewable artifact.

๐Ÿ’ก

Pro tip: Have Claude expose the module number, title, and icon as variables so generating the next 10 module thumbnails is trivial.

Digital Product / Template Thumbnail

18/30

You are a marketplace designer who builds listing thumbnails for digital products and templates. <context> I sell a digital product and need a 16:9 (or square) listing thumbnail built as one self-contained HTML file with inline CSS, previewable as an artifact, made to convert in a marketplace grid. </context> <inputs> - Product name and type: [E.G. NOTION TEMPLATE, ICON PACK, UI KIT] - Core benefit: [WHAT IT DOES FOR THE BUYER] - What is included: [COUNT OR HIGHLIGHTS] - Brand colors and aesthetic: [DESCRIBE OR HEX] - Aspect ratio: [16:9 OR 1:1] </inputs> <task> Build a listing thumbnail with the product name, a clear benefit line, an "includes" cue (e.g. "50+ icons" or "12 templates"), a stylized preview mock block representing the product, and a clean premium frame. Make the value obvious at a glance in a crowded grid. </task> <constraints> - One self-contained HTML file, requested aspect ratio, Google Fonts and CSS/SVG only. - Benefit and includes-count both legible at thumbnail size; preview mock clearly labeled. - Accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then explain how the design communicates value fast in a marketplace. </format>

Builds a converting digital-product listing thumbnail with benefit, includes-count, and preview mock as a previewable artifact.

๐Ÿ’ก

Pro tip: Give Claude the exact count of what's inside ("50+ icons"); a concrete number on the thumbnail outsells a vague "big pack".

Webinar / Workshop Promo Thumbnail

19/30

You are an event designer who builds promo thumbnails for webinars and live workshops. <context> I need a 16:9 webinar promo thumbnail built as one self-contained HTML file with inline CSS, previewable as an artifact, for use on registration pages and social. </context> <inputs> - Webinar title: [EXACT TITLE] - Date and time: [WHEN] - Host name and credential: [WHO] - The promise / takeaway: [WHAT ATTENDEES GET] - Brand colors: [HEX OR DESCRIBE] </inputs> <task> Build a webinar promo with the title as the hero, a clear date-and-time band, a host photo placeholder with name and credential, a one-line promise, and a "free" or "live" badge. Make the date unmissable and the title scannable so it drives registrations. </task> <constraints> - One self-contained HTML file, exactly 16:9, Google Fonts and CSS/SVG only. - Date and title both legible at small sizes; host placeholder clearly labeled. - Accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then explain how you made the date and value prominent for sign-ups. </format>

Creates a 16:9 webinar promo thumbnail with title, date band, and host that drives registrations as a previewable artifact.

๐Ÿ’ก

Pro tip: Tell Claude to make the date its own high-contrast band โ€” a missed date is a missed registration.

Ebook / Lead-Magnet Cover Thumbnail

20/30

You are a designer who builds ebook and lead-magnet cover thumbnails. <context> I need an ebook or lead-magnet cover thumbnail built as one self-contained HTML file with inline CSS (portrait book ratio, roughly 1240x1750), previewable as an artifact, for landing pages and ads. </context> <inputs> - Ebook / guide title: [EXACT TITLE] - Subtitle or promise: [SHORT LINE] - Topic and audience: [WHO IT HELPS] - Brand colors and style: [DESCRIBE OR HEX] - Author or brand: [NAME] </inputs> <task> Build a portrait cover that looks like a real downloadable: a bold title treatment, a supporting subtitle, a topic-fitting graphic motif or icon, and an author or brand line at the bottom. Optionally add a subtle 3D-book or page-edge effect with CSS so it reads as a tangible asset on a landing page. </task> <constraints> - One self-contained HTML file, portrait book ratio, Google Fonts and CSS/SVG only. - Title legible when shown small in a hero section; clean, premium feel. - Accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then explain the cover-design choices and how to drop it onto a landing page. </format>

Builds a portrait ebook or lead-magnet cover thumbnail with optional 3D effect as a previewable artifact.

๐Ÿ’ก

Pro tip: Ask Claude for a subtle CSS page-edge or shadow; a flat rectangle looks like clip art, a slight book depth looks downloadable.

Text-Overlay & Layout Systems

5 prompts

Bold Text-Overlay Layout

21/30

You are a typographic designer who builds high-impact text-overlay thumbnails. <context> I need a 16:9 text-driven thumbnail built as one self-contained HTML file with inline CSS, previewable as an artifact, where typography does almost all the work. </context> <inputs> - The headline phrase: [3-7 WORDS] - The single word to emphasize: [WHICH WORD POPS] - Mood: [LOUD, ELEGANT, URGENT, PLAYFUL] - Background style: [SOLID, GRADIENT, PATTERN] - Color palette: [HEX OR DESCRIBE] </inputs> <task> Build a text-overlay layout where the headline is the entire design: split it across lines for rhythm, give the emphasized word a distinct color or weight, add a supporting graphic accent (underline, highlight box, or shape), and balance it on a clean background. No photos required โ€” pure type and shape. </task> <constraints> - One self-contained HTML file, exactly 16:9, Google Fonts and CSS/SVG only. - Headline readable at small sizes; clear hierarchy with one dominant word. - Accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then explain the line-break, emphasis, and accent choices. </format>

Builds a pure-typography 16:9 text-overlay thumbnail with a single emphasized word as a previewable artifact.

๐Ÿ’ก

Pro tip: Name the one word you want to pop; Claude will make every other design choice serve that single emphasis.

Reusable Thumbnail Template System

22/30

You are a design-systems specialist who builds reusable thumbnail templates with swappable slots. <context> I need a 16:9 thumbnail template built as one self-contained HTML file with inline CSS, previewable as an artifact, designed with clearly labeled slots so I can mass-produce on-brand thumbnails. </context> <inputs> - Brand colors and fonts: [DESCRIBE OR HEX] - Fixed brand elements: [LOGO TEXT, ACCENT BAR] - Variable slots I need: [E.G. HEADLINE, BADGE, ICON, IMAGE ZONE] - Channel: [YOUTUBE / BLOG / COURSE] - Sample content to demo: [ONE EXAMPLE] </inputs> <task> Build a template with a fixed brand frame and clearly defined variable slots (headline, badge, icon, image placeholder), each driven by a CSS variable or a clearly commented section. Demo it filled with the sample content, but structure it so producing the next thumbnail means editing only the variable zone. </task> <constraints> - One self-contained HTML file, exactly 16:9, Google Fonts and CSS/SVG only. - Variables and slots clearly named and commented at the top; fixed vs variable zones obvious. - Accessible contrast; no filler beyond the demo content. </constraints> <format> Return the full HTML in an artifact, then list every variable and how to change it for the next thumbnail. </format>

Builds a reusable 16:9 thumbnail template with labeled swappable slots and CSS variables as a previewable artifact.

๐Ÿ’ก

Pro tip: Ask for a variables block at the very top of the CSS; that turns the artifact into a reusable factory, not a one-off.

SVG Thumbnail (Crisp at Any Size)

23/30

You are a vector designer who builds resolution-independent thumbnails as pure SVG. <context> I need a 16:9 thumbnail built as a single self-contained inline SVG (in an HTML wrapper), previewable as an artifact, that stays crisp at any size and is easy to edit by hand. </context> <inputs> - Headline text: [PHRASE] - Visual concept / icon idea: [WHAT TO ILLUSTRATE] - Color palette: [HEX] - Style: [FLAT, GEOMETRIC, OUTLINE] - Use case: [YOUTUBE / BLOG / SLIDE] </inputs> <task> Build a 1280x720 viewBox SVG thumbnail using vector shapes and SVG text: a clean background, a simple iconographic illustration built from paths or basic shapes, and a bold headline. Organize it with grouped, commented elements so each part is editable. Wrap it in minimal HTML so it previews. </task> <constraints> - Single inline SVG inside a minimal HTML wrapper; no raster images, no external assets except web-safe or Google Fonts. - Use a 1280x720 viewBox; readable text; grouped and commented shapes. - Accessible contrast; no filler. </constraints> <format> Return the full SVG/HTML in an artifact, then label the groups so I can recolor or edit shapes by hand. </format>

Builds a resolution-independent 16:9 SVG thumbnail with grouped, editable vector shapes as a previewable artifact.

๐Ÿ’ก

Pro tip: Because it's vector, ask Claude to group shapes with comments so you can recolor the whole thing by changing two fill values.

Brand-Consistent Thumbnail Style Guide

24/30

You are a brand designer who documents a thumbnail style as a one-page visual guide. <context> I want a thumbnail style guide built as one self-contained HTML file with inline CSS, previewable as an artifact, that defines the rules and shows example thumbnails so any creator stays on-brand. </context> <inputs> - Brand name and colors: [NAME PLUS HEX] - Fonts: [HEADLINE AND BODY] - Channel and content type: [DESCRIBE] - Do's and don'ts I care about: [E.G. ALWAYS BIG TEXT, NEVER MORE THAN TWO COLORS] - Example topics: [2-3 SAMPLE THUMBNAIL TOPICS] </inputs> <task> Build a one-page style guide showing: the color palette swatches, the type scale, the safe-area and minimum-text-size rules, a do's-and-don'ts comparison, and two or three rendered example thumbnails that follow the rules. Make it a practical reference a teammate could follow. </task> <constraints> - One self-contained HTML file, Google Fonts and CSS/SVG only. - Show real swatches, real example thumbnails (16:9 mini-frames), and concrete rules. - Accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then summarize the three rules that matter most for consistency. </format>

Builds a one-page thumbnail style guide with palette, type scale, rules, and example frames as a previewable artifact.

๐Ÿ’ก

Pro tip: List your two non-negotiable rules in the inputs; Claude will turn them into the do's-and-don'ts section everyone references.

Slide / Carousel Cover Thumbnail

25/30

You are a social designer who builds cover slides for LinkedIn and Instagram carousels. <context> I need a square (1080x1080) carousel cover slide built as one self-contained HTML file with inline CSS at 1:1, previewable as an artifact, designed to stop the scroll and make people swipe. </context> <inputs> - Carousel topic / promise: [WHAT THE SLIDES DELIVER] - Big cover headline: [HOOK PHRASE] - Number of slides or steps: [E.G. "7 STEPS"] - Brand colors and font: [DESCRIBE OR HEX] - Swipe cue style: [ARROW, "SWIPE", DOTS] </inputs> <task> Build a cover slide with a bold hook headline, a clear value or count cue, a subtle but visible swipe prompt (arrow or "swipe" label), and a small brand or handle lockup. Optimize it for the square feed where the cover decides whether anyone swipes. </task> <constraints> - One self-contained HTML file, exactly 1:1, Google Fonts and CSS/SVG only. - Hook headline dominant and mobile-legible; swipe cue clear but not loud. - Accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then explain the hook and swipe-cue decisions that drive the swipe. </format>

Builds a 1080x1080 carousel cover slide with a scroll-stopping hook and swipe cue as a previewable artifact.

๐Ÿ’ก

Pro tip: Put a visible swipe cue on the cover; carousels that don't tell people to swipe quietly lose half their reach.

Most people use 10% of Claude. Tutorials unlock the rest.

AI Academy: 300+ hands-on tutorials on Claude, ChatGPT, Midjourney, and 50+ AI tools. New tutorials added every week.

Start Your Free Trial

Niche & A/B Variant Thumbnails

5 prompts

Gaming / Stream Thumbnail

26/30

You are a thumbnail designer for gaming videos and streams. <context> I need a 1280x720 gaming thumbnail built as one self-contained HTML file with inline CSS at 16:9, previewable as an artifact, with the high-energy look gaming audiences expect. </context> <inputs> - Game and video type: [GAME, E.G. SPEEDRUN / TIER LIST / REACTION] - Overlay text: [SHORT HYPE PHRASE] - Streamer face expression: [HYPED / FOCUSED / SHOCKED] - Color energy: [NEON, DARK, VIBRANT] - Key visual element: [CHARACTER, SCORE, ITEM] </inputs> <task> Build a gaming thumbnail with a face-cam placeholder (labeled with the expression), a saturated game-scene placeholder block, a punchy outlined overlay phrase, and energetic accents (glows, speed lines, neon edges) using CSS. Push contrast and saturation while keeping the text readable. </task> <constraints> - One self-contained HTML file, exactly 16:9, Google Fonts and CSS/SVG only. - High energy but text legible at small sizes; clear focal point. - Accessible contrast on the headline plate; no filler. </constraints> <format> Return the full HTML in an artifact, then explain the energy devices and how to keep it readable in the feed. </format>

Builds a high-energy 16:9 gaming thumbnail with face cam, neon accents, and hype text as a previewable artifact.

๐Ÿ’ก

Pro tip: Tell Claude the exact reaction face; in gaming, an over-the-top expression plus neon contrast is most of the click.

Vlog / Lifestyle Thumbnail

27/30

You are a thumbnail designer for vlogs and lifestyle content. <context> I need a 1280x720 vlog thumbnail built as one self-contained HTML file with inline CSS at 16:9, previewable as an artifact, with a warm, personal, less-clickbait feel. </context> <inputs> - Vlog topic / moment: [WHAT HAPPENED] - Location or theme: [WHERE OR WHAT] - Short overlay text: [SOFT HOOK PHRASE] - Mood and palette: [WARM, AIRY, MOODY] - Creator face: [SHOWN OR NOT] </inputs> <task> Build a vlog thumbnail with a lifestyle scene placeholder block, an optional creator face placeholder, a tasteful overlay phrase in a friendly type treatment, and a warm color grade feel. Keep it inviting and authentic rather than loud, while still being scannable. </task> <constraints> - One self-contained HTML file, exactly 16:9, Google Fonts and CSS/SVG only. - Warm, legible, not garish; text readable at small sizes. - Accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then explain how you kept it authentic while still scroll-stopping. </format>

Builds a warm, authentic 16:9 vlog thumbnail with a soft hook and lifestyle scene as a previewable artifact.

๐Ÿ’ก

Pro tip: Ask Claude for a warm color grade and a softer font; vlog audiences trust thumbnails that feel personal, not screaming.

Faceless / Stock-Style Thumbnail

28/30

You are a thumbnail designer for faceless channels (no on-camera person). <context> I need a 1280x720 faceless thumbnail built as one self-contained HTML file with inline CSS at 16:9, previewable as an artifact, that works without any human face. </context> <inputs> - Video topic: [SUBJECT] - Niche: [E.G. FINANCE, TECH, MOTIVATION] - Big overlay text: [HOOK PHRASE] - Symbol or object to feature: [ICON OR THING] - Color palette: [HEX OR DESCRIBE] </inputs> <task> Build a faceless thumbnail that leans on bold typography plus a strong symbolic graphic (object, icon, or abstract motif via CSS/SVG) instead of a face. Use a dominant headline, a clear focal symbol, and a confident color scheme so it competes in the feed without a human subject. </task> <constraints> - One self-contained HTML file, exactly 16:9, Google Fonts and CSS/SVG only. - No face; the type and symbol carry the design; text legible at small sizes. - Accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then explain how you created visual interest without a face. </format>

Builds a faceless 16:9 thumbnail driven by bold type and a symbolic graphic as a previewable artifact.

๐Ÿ’ก

Pro tip: Pick one strong symbol for the topic; faceless thumbnails win on a single iconic object plus huge readable text.

Two-Variant A/B Thumbnail Pair

29/30

You are a CTR optimization designer who creates A/B thumbnail variants for testing. <context> I need two distinct thumbnail variants for the same video, built as one self-contained HTML file with inline CSS showing both 16:9 frames side by side, previewable as an artifact for an A/B test. </context> <inputs> - Video topic / hook: [WHAT IT PROMISES] - Overlay text idea: [PHRASE OR LET YOU VARY IT] - The variable to test: [E.G. FACE VS NO FACE, NUMBER VS QUESTION, COLOR] - Niche and audience: [DESCRIBE] - Brand colors: [HEX OR DESCRIBE] </inputs> <task> Build two clearly different thumbnail variants (A and B) for the same video, each in its own 16:9 frame, that deliberately differ on the single variable being tested (e.g. headline angle, color, or layout) while keeping everything else comparable. Label each as Variant A and Variant B with a one-line note on the hypothesis. </task> <constraints> - One self-contained HTML file, two 16:9 frames, Google Fonts and CSS/SVG only. - The two variants differ on ONE controlled variable; everything else stays comparable. - Both legible at small sizes; accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then state the hypothesis and what a win on each variant would tell me. </format>

Builds two controlled A/B thumbnail variants side by side with a stated hypothesis as a previewable artifact.

๐Ÿ’ก

Pro tip: Change exactly one thing between A and B; if both differ on color and text, a win won't tell you which one caused it.

Four-Way Variant Thumbnail Grid

30/30

You are a thumbnail designer who rapidly explores creative directions before committing. <context> I need four different thumbnail directions for one video, built as one self-contained HTML file with inline CSS showing a 2x2 grid of 16:9 frames, previewable as an artifact for fast comparison. </context> <inputs> - Video topic / hook: [WHAT IT PROMISES] - Core overlay phrase: [HEADLINE] - Niche and audience: [DESCRIBE] - Brand colors: [HEX OR DESCRIBE] - Directions to explore: [E.G. BOLD TYPE, FACE-LED, NUMBER-LED, MINIMAL] </inputs> <task> Build a 2x2 grid of four genuinely distinct thumbnail concepts for the same video: each frame takes a different creative angle (e.g. typographic, face-led, number-led, minimal) while staying on-brand. Caption each frame with its concept name so I can pick a direction fast. </task> <constraints> - One self-contained HTML file, four 16:9 frames in a 2x2 grid, Google Fonts and CSS/SVG only. - Each concept visually distinct, not minor reskins; all legible at small sizes. - Accessible contrast; no filler. </constraints> <format> Return the full HTML in an artifact, then briefly recommend which of the four you'd test first and why. </format>

Builds a 2x2 grid of four distinct thumbnail concepts for one video with a recommendation as a previewable artifact.

๐Ÿ’ก

Pro tip: Ask Claude to make the four directions genuinely different, then have it recommend one โ€” a fast way to break creative block.

Frequently Asked Questions

Yes. These prompts ask Claude to return a finished thumbnail as a self-contained HTML/CSS or inline SVG artifact rendered at the correct aspect ratio. You preview it instantly in the artifact panel, tweak the text or colors, and screenshot or export it as an image.
Open the rendered artifact, set your browser or screen zoom so the frame is at full size, and screenshot the 16:9 (or square) frame. For pixel-exact output, ask Claude to set the artifact to the precise dimensions (e.g. 1280x720) and take the screenshot at 100% zoom, or use the inline SVG version and export it directly.
Each prompt fixes the aspect ratio for its platform: 16:9 (1280x720) for YouTube, 1:1 for podcast covers and carousels, 1200x630 for blog and Open Graph cards, and 2:3 for Pinterest. Tell Claude the exact pixel size you need and it will lock the frame to it.
No. The prompts use clearly labeled placeholder blocks where a face or product photo would go, so the layout is complete and previewable without any image. Drop your real photo into that zone afterward, or use the faceless prompts that rely on bold type and symbols instead.
Use the template and style-guide prompts. They build a fixed brand frame with swappable slots exposed as CSS variables, so producing the next thumbnail means editing only the headline, badge, or icon. Reuse the same fonts and accent colors across every prompt to keep a channel or course library cohesive.
Yes. The A/B and four-way variant prompts produce multiple thumbnails for the same video that differ on one controlled variable, complete with a stated hypothesis. That lets you run a clean test where a winning variant actually tells you which design choice drove the higher click-through rate.

Prompts are the starting line. Tutorials are the finish.

A growing library of 300+ hands-on tutorials on ChatGPT, Claude, Midjourney, and 50+ AI tools. New tutorials added every week.

7-day free trial. Cancel anytime.