Micro Copy Intelligence logo

Micro Copy Intelligence

Community
phazurlabs
micro-copy-intelligence

1000+ microcopy and UX-writing templates by component, tone and sector: button labels, every error message and failure state, empty states, tooltips, confirmations, permission requests and notifications. Use when writing any interface string — robotic copy is the loudest slop tell.

Overview

Publisherphazurlabs
Repositorysumi
Skill namemicro-copy-intelligence
Stars
58
Forks
11
Bundled files
5
LicenseApache-2.0
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.

  • 5 bundled files

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

  • Open source

    Published by phazurlabs on GitHub. Read the source before you install it.

Installation

Install the Micro Copy Intelligence 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/phazurlabs/sumi/tree/main/skills/micro-copy-intelligence
  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/phazurlabs/sumi.git /tmp/sumi
mkdir -p .claude/skills
cp -r /tmp/sumi/skills/micro-copy-intelligence .claude/skills/micro-copy-intelligence
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Micro Copy Intelligence 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 Micro Copy Intelligence 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 Micro Copy Intelligence 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.

Micro-Copy Intelligence

Why Microcopy Matters

Microcopy is the most underestimated layer of product design. A single word on a button can change conversion by 30%. A poorly written error message causes support tickets. A delightful empty state turns a dead-end into activation. A confusing permission request tanks opt-in rates.

Most products fail at microcopy not because the words are "wrong" but because they are written by developers in a rush, never reviewed by a writer, and never tested with users. The result: vague buttons ("Submit"), hostile errors ("Invalid input"), lifeless empty states ("No data"), and permission requests that give users no reason to say yes.

This skill gives Sumi a production-grade microcopy library — 1000+ templates across every UI component, every state, every tone, and every major industry sector. Every template is copy-paste ready, with tone variants, character counts, accessibility notes, and i18n considerations.

Core Principle: Clear > Clever

The single most important rule in UX writing: clarity beats cleverness every time. Users are not reading your interface for entertainment. They are trying to accomplish a task. Every word should reduce cognitive load, not add to it.

PrincipleDescriptionExample
Clear > CleverPlain language over puns or jargon"Save changes" not "Commit your brilliance"
Specific > GenericTell them exactly what happens"Delete 3 photos" not "Delete items"
Human > RoboticWrite like a helpful person, not a machine"We couldn't find that page" not "Error 404: Resource not located"
Short > LongEvery extra word is cognitive cost"Saved" not "Your changes have been successfully saved"
Active > PassiveDirect the user with active voice"Enter your email" not "Email should be entered"
Consistent > CreativeSame action = same word everywhereAlways "Remove" or always "Delete", never both
Helpful > BlamingGuide, don't accuse"That password is too short" not "You entered an invalid password"

Voice and Tone Framework

Voice is your product's personality — it stays constant. Tone adapts to context.

Voice Spectrum (Pick Your Position)

Formal ←————————————————→ Casual
  |                           |
  Banking, Legal,         Social, Gaming,
  Enterprise              Consumer

Serious ←————————————————→ Playful
  |                           |
  Healthcare, Finance     Entertainment,
  Security                Kids, Social

Respectful ←—————————————→ Irreverent
  |                           |
  Government, Medical     Startup, Youth
  Enterprise              Consumer

Matter-of-fact ←—————————→ Enthusiastic
  |                           |
  Utilities, Tools        E-commerce,
  Productivity            Fitness, Social

Tone Adaptation by Context

The same product shifts tone based on what the user is experiencing:

ContextTone ShiftWhy
ErrorsEmpathetic, calm, helpfulUser is frustrated — don't add to it
SuccessWarm, affirming, briefUser accomplished something — celebrate proportionally
Empty statesEncouraging, instructionalUser sees nothing — guide them to value
Destructive actionsCautious, specific, clearHigh stakes — no ambiguity allowed
OnboardingWelcoming, motivatingFirst impressions — build confidence
Loading/waitingLight, reassuringUser is stuck — ease anxiety
Permission requestsTransparent, benefit-ledUser must trust you — explain why
Pricing/paymentTrustworthy, preciseMoney is involved — no vagueness
Legal/compliancePlain, clear, accessibleRequired info — make it readable

Content Hierarchy in UI

Every UI moment has three layers of information. Not all three are always needed.

LayerPurposeExample (File Upload)
PrimaryWhat to do"Upload your photo"
SecondaryWhy or context"This will be your profile picture"
TertiaryHow or constraints"JPG or PNG, max 5 MB"

Rules:

  • Primary copy is always visible
  • Secondary copy appears when context is unclear
  • Tertiary copy appears on hover, in helper text, or when errors occur
  • Mobile: reduce to primary + tertiary (space is precious)

The Microcopy Formula

For any UI element, apply this structure:

[Action Verb] + [Object] + [Benefit/Context]

Examples:

  • "Save changes" = action + object
  • "Upload photo to complete your profile" = action + object + benefit
  • "Delete this project permanently" = action + object + context
  • "Invite teammates to collaborate" = action + object + benefit

When to add benefit/context:

  • The action is unfamiliar (onboarding)
  • The action is irreversible (destructive)
  • The action requires trust (permissions, payments)
  • The action has a non-obvious outcome

Readability Scoring for UI

Target readability for UI text:

ComponentFlesch-Kincaid GradeMax WordsMax Characters
Button labelGrade 3-51-425
TooltipGrade 5-75-1580
Error headlineGrade 4-63-850
Error bodyGrade 5-710-25150
Empty state headlineGrade 4-63-850
Empty state bodyGrade 5-810-30180
Toast messageGrade 4-63-1060
Notification titleGrade 4-63-1065
Notification bodyGrade 5-710-25150
Onboarding headlineGrade 3-53-850
Onboarding bodyGrade 5-715-40250
Helper textGrade 5-75-15100
Confirmation dialog titleGrade 4-63-850
Confirmation dialog bodyGrade 5-710-30180
Permission request bodyGrade 5-715-35200
Loading messageGrade 3-52-850

Testing readability: Paste your copy into Hemingway Editor or use the Flesch-Kincaid formula. If a 12-year-old cannot understand it instantly, rewrite it.

Inclusive Language Checklist

Before shipping any microcopy:

  • Gender-neutral: "They" not "he/she". "Spouse" not "husband/wife". "Parent" not "mother/father" (unless context-specific).
  • Ability-neutral: "Select" not "Click". "View" not "See". "Enter" not "Type". Avoid "simple", "easy", "just" (what is easy for you may not be for them).
  • Age-neutral: Avoid generational slang. "Download" not "Yeet this to your device".
  • Culture-neutral: Avoid idioms that do not translate ("break a leg", "piece of cake"). No sports metaphors as primary copy.
  • Jargon-free: "Sign in" not "Authenticate". "Save" not "Persist". "Error" not "Exception".
  • No assumptions: Do not assume family structure, location, device, connection speed, or technical literacy.
  • Respectful defaults: "Preferred name" not "Nickname". "Phone number (optional)" not "Mobile number (required)".
  • Error messages blame the system, not the user: "We couldn't process that" not "You made an error".

i18n Considerations for Microcopy

Microcopy that works in English may break in translation. Plan ahead:

Text Expansion

LanguageExpansion vs. English
German+30-35%
French+15-20%
Spanish+20-25%
Italian+15-20%
Portuguese+20-30%
Russian+15-20%
Japanese-10-20% (character count, but wider glyphs)
Chinese-20-30% (character count)
Arabic+20-25% (plus RTL layout)
Korean-10-15%

Rules:

  • Buttons: Allow 40% extra width or use auto-sizing
  • Tooltips: Max 60% of container width in English to allow expansion
  • Never hard-code string widths
  • Test with German and Arabic (longest expansion + RTL)

Cultural Sensitivity

  • Humor: Puns and wordplay rarely translate. Stick to clear language.
  • Formality: German, Japanese, Korean require formal register by default. "Du" vs "Sie" in German matters enormously.
  • Color of language: "Red" means danger in the West, luck in China. Do not rely on color words alone.
  • Dates and numbers: "1/2/2026" is January 2 (US) or February 1 (most of the world). Use explicit formats.
  • Names: Not everyone has a first + last name structure. "Full name" is safer than "First name / Last name".
  • Icons with text: Thumbs-up is offensive in parts of the Middle East. Always pair icons with text labels.

String Externalization

  • Every user-facing string must be in a localization file, never hard-coded
  • Use ICU MessageFormat for plurals: {count, plural, one {# item} other {# items}}
  • Avoid string concatenation: "Welcome, " + name + "!" breaks in languages with different word order
  • Use full sentences as translation units, not fragments

Sector Tone Guides

Fintech / Banking

  • Voice: Trustworthy, precise, confident
  • Tone keywords: Secure, clear, straightforward, reliable
  • Do: "Your transfer of $500 to Alex is complete" / "Verify your identity to protect your account"
  • Don't: "Woohoo, money sent!" / "We need to make sure you're really you lol"
  • Button style: "Transfer funds", "Review statement", "Verify identity"
  • Error style: "We couldn't complete this transfer. Your account has not been charged. [Try again]"

Healthcare / Wellness

  • Voice: Empathetic, clear, reassuring
  • Tone keywords: Caring, supportive, plain, calm
  • Do: "Your appointment with Dr. Chen is confirmed for March 15 at 2:00 PM" / "Your results are ready. Tap to view."
  • Don't: "Awesome, you're booked!" / "RESULTS ARE IN!!!"
  • Button style: "Schedule appointment", "View results", "Message your care team"
  • Error style: "We couldn't schedule your appointment. Dr. Chen is available at these other times: [alternatives]"

SaaS / Productivity

  • Voice: Efficient, empowering, knowledgeable
  • Tone keywords: Smart, capable, streamlined, no-nonsense
  • Do: "Project created. Invite your team to get started." / "3 tasks due today"
  • Don't: "YAY, new project!" / "You've got stuff to do"
  • Button style: "Create project", "Invite team", "Start timer"
  • Error style: "This project name is already taken. Try a different name. [Rename]"

Social / Community

  • Voice: Warm, expressive, conversational
  • Tone keywords: Friendly, personal, inclusive, fun
  • Do: "Alex liked your photo" / "Share what's on your mind"
  • Don't: "User 4829 interacted with your content" / "Compose a status update"
  • Button style: "Share", "Like", "Comment", "Send a message"
  • Error style: "We couldn't post that right now. Your draft is saved. [Try again]"

E-commerce / Marketplace

  • Voice: Helpful, enthusiastic, trustworthy
  • Tone keywords: Exciting, confident, reassuring, action-oriented
  • Do: "Free shipping on orders over $50" / "Only 3 left in stock"
  • Don't: "Shipping fees may apply based on various factors" / "Limited inventory warning"
  • Button style: "Add to cart", "Buy now", "Apply code", "Complete purchase"
  • Error style: "This item just sold out. We'll notify you when it's back. [Get notified]"

Education / EdTech

  • Voice: Encouraging, patient, clear
  • Tone keywords: Supportive, motivating, structured, approachable
  • Do: "Great work! You completed 3 of 5 lessons." / "Pick up where you left off"
  • Don't: "CRUSHING IT" / "Resume incomplete module"
  • Button style: "Start lesson", "Continue learning", "Take quiz", "Review answers"
  • Error style: "That answer didn't match. Review the hint and try again. [Show hint]"

Developer Tools

  • Voice: Direct, precise, technical (but not jargon-heavy)
  • Tone keywords: Efficient, accurate, no-fluff, respectful of expertise
  • Do: "Build succeeded in 2.3s" / "API key created. Copy it now — it won't be shown again."
  • Don't: "Yay, your build worked!" / "Here's your secret key thingy"
  • Button style: "Deploy", "Create API key", "View logs", "Run tests"
  • Error style: "Build failed: missing dependency lodash@4.17. Run npm install lodash to fix. [View full log]"

Gaming / Entertainment

  • Voice: Energetic, immersive, personality-forward
  • Tone keywords: Exciting, playful, dramatic, rewarding
  • Do: "New high score! You beat your record by 200 points." / "Ready for the next challenge?"
  • Don't: "Score updated in database" / "Proceed to next level"
  • Button style: "Play now", "Challenge a friend", "Claim reward", "Continue quest"
  • Error style: "Connection lost. Your progress is saved. [Reconnect]"

Real Estate / Property

  • Voice: Professional, informative, aspirational
  • Tone keywords: Trustworthy, detailed, approachable, knowledgeable
  • Do: "3 new listings match your search" / "Schedule a tour of 123 Oak Street"
  • Don't: "OMG new houses!" / "Initiate property viewing request"
  • Button style: "Schedule tour", "Save listing", "Contact agent", "Get pre-approved"
  • Error style: "This listing is no longer available. Here are similar homes nearby. [View similar]"

Travel / Hospitality

  • Voice: Inspiring, helpful, warm
  • Tone keywords: Adventurous, reassuring, personal, anticipatory
  • Do: "Your trip to Tokyo is confirmed. 14 days to go!" / "Explore things to do near your hotel"
  • Don't: "Booking reference confirmed" / "Ancillary activity discovery module"
  • Button style: "Book now", "Explore destinations", "Check availability", "Add to trip"
  • Error style: "These dates just got booked. Try nearby dates — we found 5 options. [View alternatives]"

Reference Architecture

FileContentsUse When
references/button-label-templates.md200+ button label templates organized by action type: primary, destructive, navigation, toggle, social, commerce, file, communication. Each with tone variants, character counts, when-to-use guidance, anti-patterns, and i18n notes.Writing button labels. Choosing between "Save" vs "Save changes" vs "Update". Standardizing CTAs across a product.
references/error-success-messages.md200+ error and success message templates: validation, system, permission, business logic errors. Success confirmations for every common action. Each with headline + body + recovery action + ARIA announcement.Writing error messages. Creating success confirmations. Building a toast/notification system. Ensuring accessible error handling.
references/empty-state-copy.md150+ empty state copy templates: first-use, no-results, cleared, error, permission states. By component: inbox, dashboard, list, table, search, notifications, activity, calendar, kanban, files, chat. Each with headline, body, CTA, illustration suggestion, tone variants.Designing empty states. Writing first-use experiences. Creating no-results messages. Handling permission-gated blank screens.
references/onboarding-permission-copy.md150+ onboarding and permission copy templates: welcome screens, profile setup, feature tours, permission requests (notifications, location, camera, etc.), pre-permission priming, progressive disclosure, activation prompts. Each with headline, body, primary/secondary CTA, visual suggestion.Designing onboarding flows. Writing permission request screens. Creating feature tours. Building activation sequences.
references/notification-tooltip-copy.md200+ notification, tooltip, and contextual help templates: push notifications, in-app notifications, toasts, tooltips, helper text, confirmation dialogs, loading messages, changelog entries. Each with text template, character limit, tone, timing.Writing push notifications. Creating tooltip copy. Building confirmation dialogs. Writing loading states. Composing changelog entries.

How to Use This Skill

Quick Lookup

"I need a button label for [action]" — Go to button-label-templates.md, find the action type, pick the tone variant that matches your product voice.

Full Screen Copy

"I'm designing a [screen type] and need all the copy" — Combine templates from multiple reference files: buttons from button-label-templates.md, errors from error-success-messages.md, empty states from empty-state-copy.md, tooltips from notification-tooltip-copy.md.

Tone Calibration

"My product is a [sector] app with a [formal/casual] voice" — Use the Sector Tone Guides above to set your voice, then filter all templates through that tone.

Copy Audit

"Review my existing copy for quality" — Check against the principles (Clear > Clever), readability scores, inclusive language checklist, and anti-patterns in each reference file.

i18n Preparation

"We're localizing to [language]" — Check text expansion ratios, review cultural sensitivity notes, verify string externalization patterns.

Quality Checklist for Any Microcopy

Before shipping any piece of UI text, verify:

  • Is it clear on first read? (No re-reading needed)
  • Is it specific? (User knows exactly what will happen)
  • Is it actionable? (User knows what to do next)
  • Is it consistent? (Same terms used across the product)
  • Is it concise? (No unnecessary words)
  • Is it inclusive? (No gendered, ableist, or culturally specific language)
  • Is it translatable? (No idioms, puns, concatenated strings)
  • Is it accessible? (Works with screen readers, announced properly)
  • Does it match the product voice? (Tone is appropriate for context)
  • Is it tested? (Real users understood it without explanation)

Microcopy Decision Tree

User encounters UI element
  |
  +-- Is it an action? (button, link, CTA)
  |     -> button-label-templates.md
  |
  +-- Is something wrong? (error, failure, warning)
  |     -> error-success-messages.md (error section)
  |
  +-- Did something succeed? (confirmation, completion)
  |     -> error-success-messages.md (success section)
  |
  +-- Is it empty? (no content, no results, first use)
  |     -> empty-state-copy.md
  |
  +-- Is it onboarding? (first launch, feature tour, setup)
  |     -> onboarding-permission-copy.md
  |
  +-- Is it asking permission? (notifications, camera, location)
  |     -> onboarding-permission-copy.md (permissions section)
  |
  +-- Is it contextual help? (tooltip, helper text, hint)
  |     -> notification-tooltip-copy.md
  |
  +-- Is it a notification? (push, in-app, toast)
  |     -> notification-tooltip-copy.md
  |
  +-- Is it a confirmation? (delete, cancel, destructive)
  |     -> notification-tooltip-copy.md (confirmation section)

Anti-Pattern Encyclopedia (Top 25)

#Anti-PatternExampleFix
1Vague button"Submit""Create account" / "Send message"
2"Click here""Click here to learn more""Learn more about pricing"
3Blaming the user"You entered an invalid email""That doesn't look like an email address"
4ALL CAPS buttons"DELETE ACCOUNT""Delete account" (sentence case)
5Technical jargon"Error 500: Internal Server Error""Something went wrong. We're looking into it."
6Double negatives"Don't forget to not leave blank""This field is required"
7Inconsistent terms"Save" here, "Update" there, "Submit" elsewherePick one term per action type
8Wall of text in modal200-word confirmation dialog2 sentences max in dialog body
9No recovery action"Payment failed." (full stop, nothing else)"Payment failed. Check your card details and try again. [Update payment]"
10Forced enthusiasm"AMAZING! You signed up!!!""Welcome. Your account is ready."
11Guilt-tripping"No thanks, I don't want to save money""No thanks" / "Skip" / "Maybe later"
12Ambiguous "OK"Dialog with only "OK" and "Cancel""Delete photo" and "Keep photo"
13Placeholder as labelInput with only "Email" as placeholderPersistent label above + placeholder example
14Disappearing feedbackToast that vanishes in 1 secondMinimum 4 seconds + dismiss button
15Generic empty state"No items""No projects yet. Create your first project to get started. [Create project]"
16Permission without contextSystem dialog with no pre-screenPre-permission screen explaining benefit first
17Error with no specifics"Invalid" (entire message)"Password must be at least 8 characters"
18Cute over clear"Oopsie daisy! Something went wrong uwu""Something went wrong. Try refreshing the page."
19Hidden destructive actionDelete styled as primary actionRed text, confirmation dialog, clear consequences
20Loading with no contextSpinner with no message forever"Loading your dashboard..." with progress indicator
21Overloaded button"Save, publish, and notify team"Three separate actions or a split button
22Negative framing"You have 0 connections""Start connecting with people you know"
23Passive voice error"An error was encountered""We hit a snag" or "Something went wrong"
24Unexplained restriction"This action is not allowed""Only workspace admins can delete projects. Ask your admin for access."
25Untranslatable idiom"Piece of cake!" as success message"Done!" or "All set!"

Writing Microcopy: Step-by-Step Process

  1. Identify the moment. What is the user doing right now? What just happened? What emotional state are they likely in?
  2. Determine the goal. What do you need the user to understand or do next?
  3. Pick the tone. Use the tone adaptation table — match the context, not your default voice.
  4. Draft with the formula. [Action] + [Object] + [Benefit/Context] as needed.
  5. Cut ruthlessly. Remove every word that is not doing work. Read it aloud — if you stumble, simplify.
  6. Check the checklist. Inclusive? Translatable? Accessible? Consistent? Specific?
  7. Test with real users. Show it out of context — can they tell what it means? Show it in context — do they do the right thing?

Accessibility Requirements for Microcopy

  • ARIA live regions: Error messages and success confirmations must be announced to screen readers via aria-live="polite" (success) or aria-live="assertive" (errors).
  • Error association: Every error message must be programmatically associated with its input via aria-describedby.
  • Button labels: If a button uses only an icon, provide aria-label with the full text label.
  • Tooltip accessibility: Tooltips must be reachable via keyboard (focus, not just hover) and announced by screen readers.
  • Loading announcements: Announce loading start and completion to screen readers.
  • Confirmation dialogs: Focus must move to the dialog when it opens and return to the trigger when it closes.
  • Status messages: Use role="status" for non-critical updates and role="alert" for urgent messages.
  • No time pressure: If a toast auto-dismisses, the information must be available elsewhere (notification center, status area).

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 Micro Copy Intelligence AI skill do?

1000+ microcopy and UX-writing templates by component, tone and sector: button labels, every error message and failure state, empty states, tooltips, confirmations, permission requests and notifications. Use when writing any interface string — robotic copy is the loudest slop tell.

Why use Micro Copy Intelligence on TypingMind?

Because you install it once and use it with any model. Micro Copy Intelligence 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 Micro Copy Intelligence in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/phazurlabs/sumi/tree/main/skills/micro-copy-intelligence. 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 Micro Copy Intelligence?

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 Micro Copy Intelligence?

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

Is the Micro Copy Intelligence AI skill free?

Yes. It is published on GitHub by phazurlabs under the Apache-2.0 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 👇