Glassmorphism logo

Glassmorphism

Organization
iart-ai
glassmorphism

This skill should be used when the user asks to "add a glassmorphism effect", "frosted glass UI", "Apple liquid glass style", "frosted blur card", "translucent glass panel animation", "make a frosted nav bar", "build a glass modal/dialog", "animate a glass card on hover", or "add a refracting liquid-glass hero". Covers frosted translucent panels with backdrop-filter blur, edge/specular highlights, SVG liquid-glass refraction, motion on hover/scroll/enter, and accessible reduced-transparency fallbacks.

Overview

Publisheriart-ai
Repositoryweb-animation-skills
Skill nameglassmorphism
Stars
71
Forks
10
Bundled files
1
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by iart-ai on GitHub. Read the source before you install it.

Installation

Install the Glassmorphism AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/iart-ai/web-animation-skills/tree/main/skills/glassmorphism
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/iart-ai/web-animation-skills.git /tmp/web-animation-skills
mkdir -p .claude/skills
cp -r /tmp/web-animation-skills/skills/glassmorphism .claude/skills/glassmorphism
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Glassmorphism in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Glassmorphism on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Glassmorphism is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Glassmorphism (Frosted & Liquid Glass)

Translucent, frosted UI: panels you can see through, blurred so the backdrop reads as soft color, rimmed by a bright edge of light, floating on layered shadow. The 2026 "liquid glass" evolution adds refraction — the background visibly bends through the panel — plus moving specular highlights. The goal is depth and material, not just opacity.

When to use

  • Frosted cards, nav bars, sidebars, modals/dialogs, and toolbars over busy/photo backgrounds
  • Login/hero panels and overlays that should feel layered above the page
  • Apple-style "liquid glass" heroes: refracting, slowly churning surfaces with a specular sweep
  • Hover/scroll/enter motion on glass — frost deepening, highlight brightening, the panel lifting
  • Any place where content needs to read as floating glass while staying legible

The four ingredients of believable glass

Glass is never one property. Every panel needs all four, or it reads as a plain translucent box:

  1. Frost — backdrop-filter: blur() plus saturate() to keep backdrop color vivid through the blur.
  2. Translucency — a background with alpha (~0.08–0.18), never fully opaque, never fully clear.
  3. Edge of glass — a hairline border brighter than the fill, ideally a lit top-left → dim bottom-right highlight, so the rim catches light.
  4. Depth — layered box-shadow: an ambient drop shadow to float it, plus an inset top highlight and bottom shade to round the surface.
css
.glass {
  background: rgba(255, 255, 255, 0.12);
  backdrop-filter: blur(16px) saturate(160%);
  -webkit-backdrop-filter: blur(16px) saturate(160%); /* Safari needs the prefix */
  border-radius: 20px;
  border: 1px solid rgba(255, 255, 255, 0.25);
  box-shadow:
    0 8px 32px rgba(0, 0, 0, 0.25),     /* float */
    inset 0 1px 0 rgba(255, 255, 255, 0.35), /* top highlight */
    inset 0 -1px 0 rgba(0, 0, 0, 0.12);      /* bottom shade */
}

Frost tiers: light UI chrome blur(8–12px) / alpha 0.08–0.14; heavy modal/login blur(24–40px) / alpha 0.10–0.18. Dark glass: fill rgba(0,0,0,0.3), softer top highlight.

backdrop-filter only blurs content that sits behind the element. Over a flat background color it does nothing visible — glass needs a busy backdrop (image, gradient, video) to frost.

Specular edge highlight and anti-banding grain

A flat border is dead. Draw a masked 1px ring with a lit corner via ::before, and dither away gradient banding (stepped contour lines under heavy blur) with a faint feTurbulence noise overlay via ::after at opacity ≤ 0.06. Full masked-border and inline-noise recipes are in references/glass-recipes.md.

Motion: animate the glass, not the geometry

Animate the qualities of the glass — frost deepening, highlight brightening, the panel lifting — so it reads as the surface "leaning into the light."

css
.glass {
  transition: backdrop-filter .4s ease, background .4s ease,
              box-shadow .4s ease, transform .4s cubic-bezier(.16,1,.3,1);
}
.glass:hover {
  backdrop-filter: blur(24px) saturate(180%);
  background: rgba(255,255,255,.18);
  transform: translateY(-4px);
  box-shadow: 0 16px 48px rgba(0,0,0,.3), inset 0 1px 0 rgba(255,255,255,.5);
}
  • Entrance (frost resolving from clear): pair a [data-enter] transition with @starting-style { backdrop-filter: blur(0); opacity: 0; transform: translateY(16px) }.
  • Specular sweep: a skewed white gradient animated across the surface on hover or as a slow ambient loop.
  • Caveat: animating the blur radius is the most expensive thing here — the backdrop is re-blurred every frame. If hover janks, animate transform/box-shadow/background (cheap) and keep the blur delta small or snap it.

Liquid glass — SVG refraction (the 2026 look)

Bend the background through the panel like a thick lens: drive feDisplacementMap from animated feTurbulence, applied via filter (it distorts the element's own pixels) layered over a backdrop-filter frost.

html
<svg width="0" height="0" style="position:absolute"><filter id="liquid-glass">
  <feTurbulence type="fractalNoise" baseFrequency="0.008 0.012" numOctaves="2" seed="3" result="n">
    <animate attributeName="baseFrequency" dur="18s"
             values="0.008 0.012;0.012 0.008;0.008 0.012" repeatCount="indefinite"/>
  </feTurbulence>
  <feDisplacementMap in="SourceGraphic" in2="n" scale="22" xChannelSelector="R" yChannelSelector="G"/>
</filter></svg>

scale = refraction strength (8–14 subtle, 20–30 pronounced, >40 molten). Lower baseFrequency = larger, slower waves (more liquid). Keep the animate dur long (12–20s) so it breathes. These filters are heavy — use on one hero element, not a grid.

Performance — backdrop-filter is expensive

  • Each glass element re-samples and blurs everything behind it. Limit the count; a dozen on screen tanks scroll FPS.
  • Promote heavy glass to its own layer (transform: translateZ(0) or will-change: transform), and remove will-change after the animation.
  • Don't stack backdrop-filter elements over each other — nested blur re-blurs the blur (quadratic cost).
  • Big radii (>40px) over large areas are the worst case; cap radius and panel size.

Accessibility — non-negotiable

Glass is decoration; content must stay legible and the effect must be defeatable.

css
@media (prefers-reduced-transparency: reduce) {
  .glass { background: #1a1a2e; backdrop-filter: none; -webkit-backdrop-filter: none; }
  .glass::after { display: none; }            /* drop grain */
}
@media (prefers-reduced-motion: reduce) {
  .glass { transition: none; }
  .liquid { filter: none; }                   /* stop the refraction churn */
}
@supports not (backdrop-filter: blur(1px)) {
  .glass { background: rgba(30,30,50,.85); }  /* opaque-enough fallback */
}

Contrast over busy backgrounds: translucent panels over photos can drop text below WCAG. Add a semi-opaque scrim or darkening gradient under the text, optionally text-shadow: 0 1px 2px rgba(0,0,0,.4), and verify ≥ 4.5:1 body / ≥ 3:1 large against the brightest region of the backdrop.

Deliver & verify (standalone HTML)

Packaged helper (scripts/): scripts/seek-shot.sh anim.html 0 1.5 3 freezes the ?t=N harness and screenshots each moment; scripts/contact-sheet.sh sheet.png frame-*.png tiles them for one-glance review. See scripts/README.md.

For a self-contained glass demo (card, nav, modal, liquid-glass hero) the deliverable is one HTML file that opens directly in a browser — markup, CSS, and (if liquid) one inline SVG filter, no build step. A single file is the right tier for a glass surface; don't reach for a bundler. The panel must overlap a real backdrop (drop in a photo or gradient) or there is nothing to frost and the demo lies.

Output contract:

  • One .html file: a busy backdrop, your glass markup, the CSS (frost + edge highlight + grain + shadows), and any inline SVG feDisplacementMap filter.
  • A way to land on the resolved state and a frozen frame of any loop — glass verifies by both state (rest/hover/enter) and time (the sweep/refraction).
  • Include the freeze harness below so any moment can be pinned for inspection.

Seek harness — pin a deterministic state + frame. ?state=hover forces the hover look; ?t=N freezes CSS animations and the SVG refraction at N seconds, so a screenshot lands on a still, deterministic frame:

html
<script>
  const p = new URLSearchParams(location.search);
  const state = p.get("state");
  if (state) document.documentElement.dataset.state = state;   // CSS keys off :root[data-state="hover"]
  const t = p.get("t");
  if (t !== null) {
    document.getAnimations().forEach(a => { a.currentTime = parseFloat(t)*1000; a.pause(); }); // freeze CSS sweep
    const svg = document.querySelector("svg");
    svg?.pauseAnimations?.(); svg?.setCurrentTime?.(parseFloat(t));                            // freeze refraction
  }
  window.__ready = true;
</script>

Verify loop — render → set state/freeze → screenshot → check: open each meaningful state (?state=rest, ?state=hover) and a frozen loop frame (?t=2), screenshot, and confirm the glass reads as glass — blur and edge highlight visibly present, no color banding, text contrast holds over the backdrop — plus check artifacts (flat box because the backdrop is plain color, missing -webkit- prefix in Safari, refraction too strong/molten, jank from animating blur). Then load prefers-reduced-transparency and confirm the solid fallback renders legibly. Any headless tool works:

bash
npx playwright screenshot --wait-for-timeout=500 "file://$PWD/glass.html?state=hover" hover.png
npx playwright screenshot --wait-for-timeout=500 "file://$PWD/glass.html?t=2"          loop.png

To check the fallback headlessly, emulate the media feature (e.g. Playwright --color-scheme-style flags, or a ?reduced=1 query that toggles a class mirroring the media query) and screenshot that the panel is solid and readable.

Before you finish:

  1. Opens standalone — no console errors; the glass overlaps a real busy backdrop, not a flat color.
  2. Reads as glass — blur + bright edge highlight both visible, layered shadow floats it, no banding (grain present).
  3. Text contrast holds over the brightest part of the backdrop (≥ 4.5:1 body).
  4. ?state= / ?t=N freeze lands a deterministic, settled frame.
  5. prefers-reduced-transparency → solid opaque panel; prefers-reduced-motion → refraction/sweep stopped; @supports fallback for no backdrop-filter.
  6. Performance sane — glass count limited, blur not animated on a janky path, liquid refraction on one element only.

Quick reference

NeedApproach
Base frostbackdrop-filter: blur(16px) saturate(160%) + alpha bg
Edge of glassbrighter border + masked ::before lit-corner ring
Float / depthbox-shadow drop + inset top highlight & bottom shade
Kill bandingfeTurbulence noise ::after, opacity ≤ .06
Hover motiontransition transform/box-shadow/background, small blur delta
Entrance@starting-style { backdrop-filter: blur(0); opacity:0 }
Liquid refractionfeDisplacementMap ← animated feTurbulence, via filter
Specular sweepskewed white gradient animated across surface
Safarialways pair -webkit-backdrop-filter
Reduced transparencysolid opaque panel, drop grain
No backdrop-filter@supports not (...) opaque-enough fallback

Reference files

  • references/glass-recipes.md — full copy-paste recipes: base glass card with tiers, the masked specular edge ring, anti-banding grain, hover/enter/scroll motion, the specular sweep, the SVG feDisplacementMap/feTurbulence liquid-glass filter with tuning knobs, performance rules, the reduced-transparency/reduced-motion/@supports fallbacks, contrast defenses, and the standalone freeze harness.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Glassmorphism AI skill do?

This skill should be used when the user asks to "add a glassmorphism effect", "frosted glass UI", "Apple liquid glass style", "frosted blur card", "translucent glass panel animation", "make a frosted nav bar", "build a glass modal/dialog", "animate a glass card on hover", or "add a refracting liquid-glass hero". Covers frosted translucent panels with backdrop-filter blur, edge/specular highlights, SVG liquid-glass refraction, motion on hover/scroll/enter, and accessible reduced-transparency fallbacks.

Why use Glassmorphism on TypingMind?

Because you install it once and use it with any model. Glassmorphism is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Glassmorphism in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/iart-ai/web-animation-skills/tree/main/skills/glassmorphism. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Glassmorphism?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Glassmorphism?

As many as you like. As long as a model supports skills, you can use Glassmorphism with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Glassmorphism AI skill free?

Yes. It is published on GitHub by iart-ai under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇