youthumb-prompts

Original🇺🇸 English
Translated

Generate optimized prompts for YouThumb.ai YouTube thumbnails. Guided 4-step workflow: collect person name, map visual assets, describe the video, then generate 5 distinct ready-to-paste prompts. Use when the user says "thumbnail prompt", "YouThumb prompt", "generate thumbnail", "miniature YouTube", "prompt for my thumbnail", "help me with YouThumb", or when preparing YouTube thumbnail prompts.

2installs
Added on

NPX Install

npx skill4agent add mikecodeur/skills youthumb-prompts

YouThumb Prompts — Thumbnail Prompt Generator

Generate 5 optimized, ready-to-paste prompts for YouThumb.ai thumbnails.
Follow the 4 steps in order. Do not skip ahead or generate prompts before collecting all info.

Step 1 — Person

Ask the user:
Who will be on the thumbnail? Give me the name exactly as registered in YouThumb.ai.
Store the name. This is the subject referenced in every prompt.
If the user hasn't created a Person yet:
Go to YouThumb.ai → Persons → Create. Upload 3-5 clear face photos (different angles, good lighting). Name it exactly as you want to reference it.

Step 2 — Assets

Ask the user:
What visual elements should appear on the thumbnail? List each with:
  • Name — what it is (logo, screenshot, icon...)
  • Description — what it looks like or represents
Examples:
  • "Claude Code logo — purple/orange terminal icon"
  • "Screenshot of my app — dark theme with charts"
Map each asset to
@image1
,
@image2
,
@image3
etc. (max 5, in order given).
Display the mapping back:
@image1 → Claude Code logo (purple/orange terminal icon)
@image2 → VS Code logo (blue code editor icon)
⚠️ Every asset MUST be referenced as
@imageN (description)
in the prompts.
Without this, YouThumb will NOT include the image.

Step 3 — Video Description

Ask the user:
Describe your video:
  • What's the topic?
  • What's the angle/hook? (comparison, tutorial, news, opinion, reveal...)
  • What emotion should the thumbnail convey?
  • Any style preference? (dark, colorful, minimal, dramatic...)

Step 4 — Generate 5 Prompts

Using Steps 1-3, generate 5 distinct prompt variations, each 100-300 words.

Prompt Structure (all 7 blocks required in each prompt)

Every prompt must describe a complete scene including ALL of these elements:
  1. Subject — Person name, position (left/center/right), pose, body framing
  2. Expression — Face expression with micro-details (eyes, mouth, eyebrows)
  3. Assets — Every
    @imageN (description)
    with placement, size, and visual treatment
  4. Background — Color scheme, gradient, glow, particles, depth
  5. Lighting — Light source direction, intensity, color temperature, shadows
  6. Materials & Effects — Textures, reflections, distortions, glitch, glass, chrome, particles
  7. Composition — Layout, depth of field, camera angle, visual hierarchy

The 5 Variations

Each prompt must have a distinct visual identity:
#ApproachMood
1Clean & ProfessionalDark minimal, confident expression, subtle glow
2Dramatic & CinematicStrong contrast, dramatic pose, particles/effects
3Bold & ColorfulVibrant accents, energetic composition, eye-catching
4Mysterious / InsiderDark mood, secretive expression, smoke/shadows
5Pattern BreakUnusual angle, unexpected element, curiosity trigger

Reference Libraries

Use these to build rich, specific prompts. Mix and match across prompts.

Expressions

ExpressionWhen to usePrompt keywords
Confident smirkDefault for most
confident smirk, knowing look, slight head tilt, one eyebrow slightly raised
ImpressedDiscovery, review
impressed but composed, raised eyebrow, subtle smirk, hint of admiration
Serious/focusedDeep tech, security
focused expression, intense gaze, determined look, sharp eyes
Finger on lipsSecret, insider
finger on lips, knowing smirk, secretive look
PointingCall to action
pointing at [element], confident expression, direct eye contact
Arms crossedAuthority, opinion
arms crossed, confident stance, authoritative look, slight smirk
Subtle surpriseGenuine wow (rare)
subtly surprised, eyebrows slightly raised, mouth slightly open
⚠️ Rule: max 1 out of 5 prompts can use "surprised". Default = confident/serious/impressed.

Backgrounds

MoodColorsKeywords
Tech/darkBlack + orange/teal
clean dark background, subtle orange glow, soft ambient light
DramaticPurple + electric blue
dark cinematic background, deep purple gradient, electric blue rim light
MinimalPure black + one accent
pure black background, single [color] light source, ultra minimal
WarmDark + amber/gold
warm dark background, golden ambient light, subtle bokeh
ColdDark blue/steel
cold steel blue background, desaturated tones, sharp shadows
EnergeticVibrant gradient
vibrant [color] to [color] gradient, dynamic energy, subtle particles
⚠️ Never: busy patterns, geometric fractals, lines everywhere, more than 2 accent colors.

Materials & Effects

EffectKeywordsBest for
Glass
frosted glass overlay, glass morphism, transparent panels
Modern tech
Chrome
chrome reflections, metallic surface, brushed steel
Premium/authority
Neon
neon glow effect, light trails, luminous edges
AI/futuristic
Holographic
holographic display, iridescent shimmer, prismatic light
Innovation
Particles
floating particles, dust motes in light, subtle sparkles
Cinematic
Glitch
subtle glitch effect, digital distortion, scan lines
Security/hacking
Smoke
wisps of smoke, atmospheric fog, volumetric haze
Mystery
Fire/embers
floating embers, fire particles, warm sparks
Hot takes
Bokeh
bokeh light orbs, out-of-focus lights, dreamy depth
Personal/vlog
Light rays
volumetric light rays, god rays, dramatic light beams
Announcements

Text & Typography

ElementKeywords
Bold text
bold white text "[TEXT]" in large sans-serif font, high contrast
Glowing text
glowing [color] text "[TEXT]", neon style lettering
3D text
3D extruded text "[TEXT]", metallic finish, dramatic shadow
No text
no text, no typography, clean visual only
Code snippet
floating code snippet, monospace font, syntax highlighted

Composition & Camera

StyleKeywords
Classic YouTube
medium shot, person on left third, element on right, shallow depth of field
Centered power
centered subject, symmetrical composition, straight-on camera angle
Dynamic angle
slight low angle, dynamic perspective, subject looking down at camera
Split frame
split frame, person on one side, visual element on other, strong contrast
Close-up
close-up portrait, tight crop, intense eye contact, blurred background

Output Format

Present each prompt like this:
━━━ PROMPT 1 — Clean & Professional ━━━

[Full prompt text ready to paste into YouThumb.ai]

Expression: confident | Background: dark + teal glow | Assets: @image1 center, @image2 floating

━━━ PROMPT 2 — Dramatic & Cinematic ━━━

[Full prompt text...]

Expression: serious | Background: purple + blue rim | Assets: @image1 large left

...etc for all 5
Add a one-line summary under each prompt (expression + background + asset placement).

Prompt Writing Rules

✅ Always

  • Reference EVERY asset with
    @imageN (description)
    — or it won't appear
  • Specify person position (left/center/right) and framing (medium/close-up)
  • Include lighting direction and color
  • Describe the scene as a whole, not a list of disconnected elements
  • Use cinematic/photographic language
  • Keep each prompt between 100-300 words

❌ Never

  • Don't write "generate a thumbnail of..." — describe the scene directly
  • Don't leave assets unplaced — every
    @imageN
    needs position + visual treatment
  • Don't use more than 2 accent colors per prompt
  • Don't describe cluttered backgrounds (lines, geometric patterns, fractals)
  • Don't default to "mouth wide open surprised" expression
  • Don't forget lighting — it makes or breaks the mood

YouThumb.ai Advanced Options Reference

When pasting prompts into YouThumb.ai, users can also set:
OptionValuesDefault
Face Expressionneutral, happy, surprised, excited, serious, confident, keep-originalconfident
Text Positiontop, center, bottom, none, keep-originalnone
Variations1-41 (each costs credits)
Clothing Stylecasual, professional, sporty, elegant, streetwear, keep-originalkeep-original
Negative PromptMax 500 chars — exclude unwanted elements
Background Blur0-1000
Face Enhancementsubtle, normal, enhanced, keep-originalnormal

Detailed Examples

For full prompt examples across 6 video types (tutorial, comparison, news, opinion, product review, security), read references/prompt-examples.md.
Load it when you need inspiration for a specific video category.

Example Walkthrough

Step 1 — Person: "Alex Dev"
Step 2 — Assets:
  • @image1 → React logo (blue atom icon)
  • @image2 → Claude Code logo (purple/orange terminal icon)
Step 3 — Video: "Comparing React development with and without AI. Angle: productivity boost. Mood: impressive, tech-forward."
One of the 5 generated prompts:
Alex Dev on the left third of frame, medium shot from waist up, confident smirk with one eyebrow slightly raised, arms crossed. @image1 (React blue atom logo) floating at mid-right, softly glowing with a blue aura. @image2 (Claude Code purple terminal icon) hovering below and behind, emitting subtle orange light trails. Clean dark background with a deep navy-to-black gradient, single teal light source from upper left casting soft directional shadows. Subtle floating particles catching the light. Chrome-like reflective floor adding depth. Shallow depth of field with background elements slightly soft. No text.
Expression: confident | Background: dark navy + teal | Assets: @image1 mid-right glowing, @image2 behind lower