Humanize Ui logo

Humanize Ui

Community
umitkaanusta
humanize-ui

Give your coding agent some design taste. Humanize your UI with proven components.

Overview

Publisherumitkaanusta
Repositoryhumanize-ui
Skill namehumanize-ui
Stars
66
Forks
6
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 umitkaanusta on GitHub. Read the source before you install it.

Installation

Install the Humanize Ui 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/umitkaanusta/humanize-ui/tree/main/skills/humanize-ui
  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/umitkaanusta/humanize-ui.git /tmp/humanize-ui
mkdir -p .claude/skills
cp -r /tmp/humanize-ui/skills/humanize-ui .claude/skills/humanize-ui
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Humanize Ui 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 Humanize Ui 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 Humanize Ui 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.

humanize-ui

Use this skill when improving or building frontend UI where the result should feel more polished than a generic AI-generated interface.

The operating model is: inspect the product first, make deliberate project-specific design choices second, scout strong human-designed references third, then adapt the best-fitting pieces into the existing frontend foundation. Treat component libraries like LEGO, not like a theme you paste over the app.

The skill should feel seamless in use. Do not stop the task to ask for design direction unless the missing choice would materially change the result. When the brief is underspecified, state the concrete decisions you are making and proceed.

Workflow

  1. Inspect the existing frontend before looking for components.

    • Identify the framework, routing model, component structure, styling system, design tokens, typography, spacing scale, colors, radii, shadows, icon set, dependencies, accessibility conventions, and responsive patterns.
    • Reuse existing primitives and local component APIs wherever sensible.
    • Do not introduce a new component library, animation package, or styling approach unless the benefit is clear and proportionate.
  2. Define the UI problem precisely.

    • Determine whether the task needs a layout pattern, page section, form, navigation element, command surface, table, dashboard module, AI interaction, empty/loading/error state, animation, transition, or interaction polish.
    • Identify what currently feels weak: hierarchy, density, contrast, alignment, affordance, state handling, motion, copy, or mobile behavior.
  3. Establish a project-specific design direction.

    • Make four explicit choices before generating UI: reference direction, color decision, type decision, and layout intent.
    • If the product already has brand or design tokens, use those choices. If it does not, pick the closest archetype in references/design-briefs.md and adapt it to the product's audience and workflow. State the direction plainly and commit to one; avoid empty directions like "modern and clean" and do not blend two briefs.
    • The layout should come from the user's real workflow, not from a default page skeleton. For example, incident software should foreground alert routing, service ownership, timelines, status updates, or monitoring context; a local repair shop should foreground booking, service menu, location, turnaround, and trust.
    • Litmus test before building: a user arriving on this screen should land inside their real task. If the screen mainly signals "this is a nicely designed page," the direction is still generic and should be redone around the workflow.
  4. Use a small visual vocabulary.

    • Pick one or two strong visual motifs and execute them consistently instead of stacking half-ideas.
    • Be deliberate about color, type, radius, borders, elevation, iconography, illustration, data display, and motion. Avoid adding gradients, glow, cards, blobs, glass, shadows, and animations just because they are available.
    • Do not replace one default look with another. Swapping AI-purple gradients for warm cream, sage green, and a refined serif is still a default if the project did not ask for it.
  5. Scout curated sources when references would improve the result.

    • Search or browse the relevant source sites for current components, code, and behavior. Prefer the actual implementation over imitating a screenshot.
    • Look for several plausible candidates when the choice affects the product experience.
    • If a source exposes install commands or registry items, inspect what it adds before integrating it.
  6. Evaluate candidates before using one.

    • Consider visual fit, UX fit, stack compatibility, dependency cost, accessibility, responsive behavior, implementation complexity, maintainability, and how easily the component can be adapted to the product's visual language.
    • Prefer one strong foundation over stitching together several unrelated aesthetics.
    • Reject components whose personality, motion, density, or dependency footprint would make the product feel inconsistent.
  7. Adapt the selected pattern into the existing design language.

    • Match local typography, spacing, colors, radius, shadows, iconography, state management, component APIs, and accessibility patterns.
    • Keep the useful layout or interaction idea while removing irrelevant styling, decorative effects, demo content, and foreign abstractions.
    • The result should look native to the project, not transplanted from another library.
  8. Polish and verify.

    • Render the result and review the actual screenshot at desktop and mobile widths. Verify against the render, not from the code alone.
    • Check hierarchy, spacing, density, alignment, hover/focus/active states, keyboard behavior, reduced-motion behavior, mobile layout, empty/loading/error states, and accessibility.
    • Audit for AI-built visual tells before finishing; treat this as a hard gate, not a soft preference. Any tell the product did not explicitly ask for gets fixed, not shipped. Untouched shadcn or Tailwind defaults, purple or blue gradient defaults, gradient hero text, unprompted neon glow, emoji-as-icons, excessive pill rounding, generic Inter or Geist by reflex, cream-plus-serif-plus-sage by reflex, and the centered hero plus three-card feature layout.
    • A flagged choice is acceptable when it is clearly intentional for the product. The problem is the unspecified default, not the color, typeface, framework, or component by itself.
    • Preserve existing functionality unless the user explicitly asks for behavior changes.
    • Improve the smallest appropriate surface area.

Curated Sources

Use these as the starting set. Search them actively when they match the task, and keep any new suggestions rare and high-quality.

  • https://beautifului.dev - Best for AI-native product surfaces: thinking states, streaming text, approval cards, tool chips, task rows, prompt bars, context cards, diff tables, records tables, search, side navigation, and other agent/workflow primitives.
  • https://beui.dev - Best for animated React/Next.js components built with Motion and Tailwind, especially command palettes, drawers, tabs, toasts, inputs, selects, docks, dynamic islands, OTP inputs, swipeable lists, theme toggles, and other interaction-heavy primitives.
  • https://rareui.com - Best for distinctive animated React components that should stand out, such as playful input states, reveal effects, visual widgets, and shadcn-installable one-file components. Use sparingly in products that can support a more expressive moment.
  • https://transitions.dev - Best for focused UI motion: dropdowns, modals, accordions, number changes, notification badges, skeleton reveals, success states, page transitions, tooltip motion, and small copy-paste CSS or React transitions.
  • https://ui.shadcn.com - Best as a customizable foundation for modern React interfaces: accessible primitives, app components, forms, tables, command menus, sidebars, dialogs, charts, blocks, and the wider registry ecosystem. Review generated or registry code before adopting it.

Integration Rules

  • Reference first, adapt second, invent only where necessary.
  • Do not blindly copy a component just because it looks good in isolation.
  • Make specific choices before styling. The work should look like it belongs to this product, this audience, and this workflow.
  • Do not create a Frankenstein UI assembled from mismatched libraries, clashing radii, incompatible shadows, random gradients, inconsistent motion, or competing interaction models.
  • Avoid stereotypical AI design flourishes unless they genuinely fit the product: oversized cards, gratuitous gradients, glassmorphism, decorative blobs, excessive rounding, dramatic shadows, and motion for its own sake.
  • Avoid the newer tasteful default too: warm cream backgrounds, sage or forest accents, refined serif display type, and soft editorial cards are not automatically more human. Use them only when they match the product's actual direction.
  • When a required page shape includes common sections such as hero, benefits, pricing, and CTA, vary the composition around the domain's workflow so it does not collapse into the same centered hero and three-card template.
  • Favor coherence over flair. A quiet, well-integrated component is usually better than a flashy imported one.
  • Apply any decorative treatment (grain, noise, gradient, blur, a distinctive border or shadow style) across the whole surface or not at all. One textured card on an otherwise flat page reads as a bug, not a highlight.
  • Use realistic content and representative data in data-heavy modules. Charts, figures, and proportions should be computed from real numbers, not approximated into decorative shapes, and never invent metrics that imply a precision the product does not have.
  • Respect licenses, provenance, and attribution requirements for copied code.
  • Before adding packages or running install commands, understand the dependency impact and follow the user's approval model.

Motion

The resting state is the design. Motion only controls how elements arrive and how they respond, and every animated element must end in its correct static position and stay usable if the animation is skipped.

  • Animate transform and opacity. Avoid animating layout properties (width, height, top, margin) that force reflow.
  • UI micro-transitions such as hover, open/close, and enter run roughly 120-240ms. Reserve slower motion for large surfaces like route changes.
  • Use one easing curve for most transitions rather than a different feel per component. A decisive ease-out such as cubic-bezier(.2, .8, .2, 1) is a safe default.
  • Stagger siblings by 40-120ms for sequence. Comparable elements (list rows, bars, cards in a set) share identical duration and easing; vary only the delay, never the speed.
  • Honor prefers-reduced-motion: drop movement and long fades, keep instant state changes.

Examples

If the user says, "Make this pricing page look better," inspect the current pricing page, identify the weak layout or hierarchy, scout pricing sections or card patterns, choose a compatible reference, then adapt spacing, typography, states, and responsive behavior to the app.

If the user says, "Add a command palette," search for high-quality command palette implementations or shadcn-compatible patterns, pick one that matches the stack and accessibility needs, inspect the code, then adapt it to local commands, styling, keyboard handling, and state management.

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 Humanize Ui AI skill do?

Give your coding agent some design taste. Humanize your UI with proven components.

Why use Humanize Ui on TypingMind?

Because you install it once and use it with any model. Humanize Ui 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 Humanize Ui in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/umitkaanusta/humanize-ui/tree/main/skills/humanize-ui. 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 Humanize Ui?

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 Humanize Ui?

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

Is the Humanize Ui AI skill free?

Yes. It is published on GitHub by umitkaanusta 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 👇