Design Glossary
Every design term explained with visual examples. Not just definitions, but the context that makes them useful.
1073 terms
0-9
60-30-10 Rule
A color distribution formula where 60% of the design uses the dominant color, 30% uses a secondary color, and 10% uses an accent color. The standard ratio for balanced visual design.
8-Point Grid
The 8-point grid constrains every logo dimension to multiples of eight pixels so strokes, radii, padding, and proportions stay consistent across any scale or handoff.
A
Above The Fold
Above the fold is the first screen of a landing page visible before scrolling. It must answer what the product is, who it is for, and why it matters in under five seconds or the rest of the page dies unseen.
Abstract Mark
An abstract mark is a geometric shape that represents a brand without showing any literal object or scene. It starts completely meaningless and only gains power after the company invests years and millions teaching the market what it stands for.
Abstract Name
An abstract name takes an existing word and assigns it to a brand in a context that shares zero connection to the word's original meaning.
Accent Calibration
Accent calibration lifts lightness and drops chroma on brand colors for dark mode while locking hue in place so the color stays recognizable without haloing or burning retinas at night.
Accent Color
The color role in a design system reserved for brand identity, primary actions, and key signals, used with discipline so it holds its meaning.
Accent Desaturation
Accent desaturation reduces saturation of UI accents by 10 to 20 percent in dark mode. This prevents brand colors from glowing harshly against dark surfaces while maintaining recognizability and improving legibility.
Accent Tier
The accent tier isolates brand and action colors inside a role-based token system so they stay consistent across states, themes, and surfaces without relying on outdated percentage rules.
Accessible Authentication
Accessible authentication requires every login, signup, or recovery flow to offer at least one method that does not rely on users remembering passwords, solving puzzles, or performing cognitive tests. WCAG 2.2 criteria 3.3.8 (AA) and 3.3.9 (AAA) make this mandatory so people with memory issues, dyslexia, or cognitive disabilities are not locked out.
Activation Surface
The activation surface is the empty state a new user hits before any data exists. This screen, not the dashboard, decides if they activate or bounce because it is the only product they have actually used.
Active Negative Space
Active negative space is the intentional carving of voids that create secondary forms, hierarchy, and meaning. It turns absence into architecture instead of treating it as leftover canvas.
Adaptive Reasoning
Claude's built-in ability to scale its thinking depth per step inside whatever effort level you set, thinking harder on hard parts and skating through easy ones.
Adaptive UI
Adaptive UI is software that reshapes its own structure, navigation, and layouts to match each user's distinct workflow and data model instead of forcing every user into one rigid schema.
Agency By Default
Agency by default is the posture where the product acts on user intent immediately, shows exactly what it did, and offers a clear undo path instead of nagging with confirmation dialogs at every step.
Agent Brief
An agent brief is the four-part document that turns vague AI wishes into reliable multi-step agents by specifying a crisp goal, the exact tools it can call, the non-negotiable rules, and the precise stop condition that ends the loop.
Agent Handoff
Agent handoff is the structured state dump that transfers task context from one agent to a human or another agent without context loss. It packages the reframed goal, plan status, current world state, decisions, artifacts, and next steps into a scannable artifact the receiver can act on in under thirty seconds.
Agent Loop
The agent loop is the visible, interruptible cycle of plan, retrieve, edit, execute, verify, and iterate that lets AI coding tools take a high-level goal and drive it end-to-end across multiple files without constant babysitting.
Agent Memory
Anything an AI product remembers about a user across sessions and then uses to change its future behavior. Storage without behavior change is just a database.
Agent Mode
Agent mode is Cursor's autonomous operation that chains multiple steps, writes to files, installs packages, runs terminal commands, and completes complex tasks without constant prompting.
Agent Turn
A complete, billable interaction cycle with an AI agent, characterized by the full resend of the entire working context, not just the visible output.
Agent-Friendly UI
Agent-friendly UI is interface design that uses semantic markup, clear labels, predictable patterns, and strong visual hierarchy so AI agents can reliably read and act on it from screenshots.
Agentic Editor
An AI code editor that accepts a high-level goal and autonomously plans steps, edits multiple files, runs tests, and completes the task with minimal real-time human input.
Agentic First Run
Agentic first run delivers a complete working artifact from one user instruction so the output itself teaches the entire product.
Aha Experience
The Aha Experience is the neurochemical brain snap that occurs when negative space suddenly reveals a second brand truth, releasing dopamine that glues the mark into long-term memory.
AI Agent
An AI agent is a long-running model that reads your full repo, makes its own decisions about which files to edit, runs tests, opens PRs, and talks back when it gets confused instead of waiting for line-by-line instructions.
AI Agent UI
AI Agent UI is the control surface for autonomous AI workers that take a goal, generate an editable plan, execute tools, and report live progress while giving users clear override and recovery options. It treats the agent as a supervised employee rather than a chat partner.
AI as Interface
AI as Interface treats the AI as the main surface users manipulate rather than a supporting feature. The output becomes the canvas, forcing designers to solve for trust, agency and mixed initiative in the same space.
AI Bolted On
AI bolted on is the 2024 anti-pattern of gluing chat sidebars and sparkle buttons onto existing products instead of rebuilding primary workflows around the model. The product works fine without the AI which makes the model optional and instantly ignorable.
AI Collaborator
An MCP-connected AI that reads your live Figma files, GitHub repos, Notion briefs, and local assets then acts inside those exact tools instead of waiting for you to paste screenshots and explain what it is looking at.
AI Design Tool
An AI design tool is a design tool with a generative or agentic model at its core, producing and editing designs or code from prompts rather than pure manual manipulation. The model is the surface you work on, not a panel bolted onto the side. Figma Weave, Google Stitch, and Vercel v0 are the current shape of it.
AI Design Workflow
An AI design workflow is a six-stage process from research to ship where AI acts as a first-class participant handling volume while humans own judgment and enforce review gates at every boundary.
AI Disclosure
The professional practice of transparently informing clients where AI tools were used in the design process, framing it as a part of craft rather than a hidden shortcut.
AI Eval
AI Eval is a structured test that scores AI design output against fixed criteria so you measure quality instead of arguing about it.
AI File
The AI file is Adobe Illustrator's native working document that holds your complete editable logo with live text, layers, global swatches, symbols, and every non-destructive decision.
AI Latency
AI latency is the delay between user input and AI output but the metric that matters is perceived speed not raw milliseconds. It is a design problem solved by streaming tokens, optimistic commits, progressive skeletons, visible reasoning, and background agents that turn wait time into usable feedback.
AI Median Pull
AI Median Pull is the statistical gravity that yanks every generative output toward the safest average in its training data, delivering Inter on cream, 12-pixel radii, flat orbs, and Stripe-Substack copy on repeat.
AI Onboarding
AI Onboarding builds a working mental model of a non-deterministic tool in the first sixty seconds by nailing four jobs: capability bound, interaction model, success state, and first prompt.
AI Orchestrator
The designer who owns end-to-end AI workflows for product features by building prompt systems, authoring Claude Skills, wiring Figma MCP pipelines, running evals, and routing output to ship measurable interfaces at scale.
AI Readable Structure
AI readable structure is semantic HTML built so machines parse your intent as cleanly as humans see your layout. In 2026 it demands one H1, logical heading flow, real section elements instead of div soup, schema.org markup that matches visible content, and alt text that describes facts instead of vibes.
AI Session
A single continuous conversation thread with an AI model, from the first message to the last. Each session has its own context window that resets when a new session starts.
AI Slop
Raw, uncurated AI-generated work shipped directly from prompt to deliverable without human judgment or refinement.
AI Token
The basic unit of text that AI language models process. Roughly 0.75 words per token in English, though the ratio varies by language and content type.
AI Voice Drift
AI voice drift is the slow erosion of a product's distinct personality when AI-generated copy, visuals, and interactions converge on generic corporate patterns pulled from training data averages.
AI Workflow Consulting
The premium engagement where a solo designer sells their own documented AI pipeline, Skill packs, prompt libraries, and integrations to clients who want the same speed and output.
AI-Assisted Work
Design work where AI tools are used as a faster pencil to generate initial directions or options, but the final output is rebuilt, refined, and owned by a human designer.
AI-Augmented Design
AI-augmented design folds large language models and custom tooling into the workflow to ship bigger systems instead of simply doing the old work faster.
AI-native
A design or system built to be composed by an AI model at request time, not assembled by hand at build time.
Alerts In Context
Alerts in context push the exact number, the one-sentence story behind it, and the action buttons directly into the user's native channel so they can resolve it without opening a dashboard.
Alias Chain
The strict three-step reference path where component tokens point to semantic tokens that point to primitive tokens, enabling theme changes without touching the components themselves.
Aligned Brief
The signed one-page document that locks audience, positioning, three brand principles, and core purpose before any pixel gets touched.
Ambient AI
Ambient AI is the pattern where the model watches the user's work without a persistent UI of its own and surfaces small, useful assists exactly when they are needed then disappears again.
Ambient Bed
The ambient bed is the seamless 60-to-90-second looping track that carries your brand's sonic DNA into background environments at the edge of attention.
Analogous
A color scheme using three to four hues that sit side by side on the color wheel, creating a cohesive and smooth visual transition.
Anchor Cell
The anchor cell is the dominant largest cell in a bento grid that earns its size from content importance and sets the visual hierarchy for all supporting cells. It typically occupies twice the area of the next largest cell and holds the strongest message or visual on the section.
Anchor Color
The single hue that carries brand recognition across every surface. It gets picked twice, once for emotional direction and once for technical fit after the neutral scale, states, and dark mode exist. Real systems ship five to nine anchor steps, never one lonely hex.
Animated Headline
One word or short phrase cycles inside an otherwise frozen hero headline while every surrounding element stays dead still. The isolated motion creates contrast that pulls the eye and lets SaaS products communicate multiple crisp use cases without turning the hero into a run-on sentence or noisy animation reel.
Anti-Dashboard
A focused interface that answers one specific question, surfaces one priority list, and ties every visible number to a clear next action while hiding everything else.
Anti-Portfolio
The anti-portfolio is a compact stack of one shipped product, one production component library, three decision logs, one coded motion demo, and public artifacts that prove shipping speed and system ownership. It replaced polished mockup sites when Linear, Vercel, and Anthropic started hiring design engineers in 2026.
APCA
The Accessible Perceptual Contrast Algorithm, a perceptual contrast formula designed to replace the WCAG 2 luminance math in the WCAG 3 draft, accounting for font size, weight, and polarity.
Applied Examples
Side-by-side rewrites of real sentences from your live copy that show the weak corporate default next to the tight brand version. They deliver instant patterns writers and AI can copy without guessing.
Approval Beat
The deliberate review pause before an AI executes any consequential action. It surfaces the exact diff, email, API payload, or plan so the user can read, edit, or reject it before anything lands.
Approval Criteria
Approval criteria are the specific written conditions a client commits to before you show brand identity work. They replace we will know it when we see it with clear measurable standards that keep presentations from becoming taste contests.
Approval Gated Tool Call
An approval gated tool call pauses the AI before any destructive or irreversible action, displays the exact plan in plain text with visible consequences, and requires the user to approve, modify, or cancel it inline inside the prompt surface.
Archetype Commitment
Archetype commitment is the discipline of selecting one brand archetype and ruthlessly aligning every downstream decision to it so the promise stays unmistakable. It turns the archetype from a workshop label into the filter that kills anything off brief.
Architecture Decision Record
Short dated records that capture one design or technical choice the real alternatives the binding constraints the reasoning the owner and the measured outcome.
Art Director
The designer who directs AI agents like Claude Code as tireless juniors, crafting CLAUDE.md briefs that encode taste and standards then reviewing every diff with ruthless precision while owning all subjective judgment.
Artisan and Craft Brand
A brand that emphasizes handmade quality, unique craftsmanship, and often the imperfect, human touch in its products and services.
Audio Mnemonic
An audio mnemonic is a short, engineered sound cue built for instant recognition and recall. It is the memory hook a brand reuses across every audio touchpoint, from the app-open sound to the checkout chime, and the sonic logo is just one place it lives.
Audit Phase
The stage after the AI shipping gold rush where founders shift from generating fast to systematically examining and strengthening the security, brand, UX, and foundation of what they built.
Audit Trail
An audit trail is the transparent log that shows exactly which memory cards, user facts, behavior signals, and scoped context shaped every agent response.
Auto-Compaction
Auto-compaction is Claude Code's unskippable safety net, automatically summarizing your conversation when the context window fills up. It is the system's last-ditch effort to keep the conversation flowing, whether you like it or not.
B
Background Agent
Background agents run complex AI tasks in side panels so users never break flow. The agent shows its plan, streams progress, edits the main canvas live, and lets the user keep working in the primary interface.
Barefoot Developer
A barefoot developer is a non-engineer who ships real software using AI tools and modern platforms. Maggie Appleton coined the term for designers, teachers, researchers and domain experts who build bespoke tools for themselves and a handful of specific people instead of chasing markets.
Base Font Size
The single root value, usually 16px set in rems, that every other size in your modular type scale multiplies or divides from using one fixed ratio.
Baseline Grid
A baseline grid is the invisible set of horizontal lines that every line of text sits on to enforce consistent vertical rhythm across an entire layout.
Bento Grid
A layout pattern of unequal rectangular cells arranged in a unified composition, replacing the three-column feature row for modern product and marketing pages.
Bespoke Software
Bespoke software is code built for one exact situation and a known handful of users instead of a generic market. The maker knows every user by name and encodes their specific quirks with zero compromise.
Bevel Survivor
A bevel survivor is a logo still carrying 3D bevels, drop shadows, inner glows, and plastic gradients from the 2008 skeuomorphic era that fails single-color, small-scale, and contemporary relevance tests.
Blanding
The 2015-2022 flattening of legacy brands into near-identical sans-serif minimalism that turned Burberry, Saint Laurent, Pinterest and Warner Bros into visual twins.
Bloated Design Case Study
The traditional, excessively long portfolio presentation that prioritizes process documentation over proof of a working product.
Bolted-on Social
Bolted-on social is the failed pattern of adding share buttons or modals as final exit steps instead of designing sharing into the core product flow from the beginning.
Bottom Tab Bar
Bottom tab bars fix three to five primary destinations in the thumb zone at the bottom of mobile screens so users tap without looking or stretching. The pattern became the default for any daily use app by 2026 because ergonomics beat every other consideration.
Branch Debt
Branch debt is the accumulated weight of abandoned conversation paths, rejected ideas, and random topic switches that an AI must keep dragging inside its context window.
Branching Prompt
A pattern where regenerating an AI response creates a new saved fork in conversation history instead of overwriting the previous output, preserving exploration paths.
Brand Archetype
A brand archetype is the single dominant personality you plant in the customer's head that dictates every visual choice, especially the typeface that must carry that position across thousands of annual impressions.
Brand Architecture
The organizational structure of brands, sub-brands, and products within a company portfolio. Defines how each entity relates to the parent brand visually and strategically.
Brand Audit
A paid two-to-five-thousand-dollar diagnostic that deconstructs a client's visual identity, messaging, hierarchy, and competitive positioning then returns a crisp Notion report with annotated fixes, type pairings, and Claude prompt packs ready to ship.
Brand Belief
Brand belief is the single unshakable conviction a company exists to prove. It sits behind every visual choice, every product decision, and every customer touchpoint, turning a logo into a signal people actually care about.
Brand Book
A brand book is the living system of tokens, rules, components, prompts and examples that keeps every asset on brand whether produced by a human or a machine.
Brand Consistency
The discipline of expressing a brand identity the same way across every format, platform, and interaction.
Brand Debt
The accumulating visual and experiential inconsistencies that appear when AI tools generate logos, colors, typography, and components across separate sessions with no shared system.
Brand Diagnostic
A decision tree that maps failing business signals to the exact layer broken in the three-layer model of identity, image, or reputation so you fix the right thing instead of defaulting to another logo refresh.
Brand Dialect
The single flattened aesthetic shared by most tech and consumer brands in 2026 built from Inter or GT Walsheim, monochrome palettes with one accent, 8-16px corner radius, flat illustrations and warm-but-professional voice.
Brand Drift
Brand drift is the slow erosion of visual and verbal consistency that happens when AI generates assets faster than humans can correct them.
Brand Editor
The brand editor governs AI-first brand systems by owning tokens, prompt packs, and evals instead of crafting individual assets. This role reviews automated scores, tunes constraints, and keeps ten thousand daily outputs consistent where classic brand directors would drown.
Brand Enemy
A brand enemy is the specific force or condition your brand exists to fight. It turns vague positioning into committed work by giving your archetype a clear antagonist to push against in every decision.
Brand Equity
The commercial value derived from consumer perception of a brand, beyond the functional value of its products or services.
Brand Eval
Brand eval is the automated test layer that scores every AI-generated asset against your token spec, voice rubric, and layout rules then kicks back failures before they ship.
Brand Framework
A living set of principles that guides every design and copy decision so the brand feels the same without looking the same across every possible context.
Brand Governance
Brand governance is the operating system for your brand rules. It defines who owns the system, how changes get proposed approved and versioned, and where the single source of truth lives so tokens voice motion and AI prompts stay consistent when humans and machines both ship assets.
Brand Guidelines
The rulebook that defines how a brand identity should be applied across every format, platform, and context.
Brand Identity
The complete visual and verbal system that makes a brand recognizable, consistent, and impossible to confuse with anyone else.
Brand Image
The current perception a person holds about your brand based on every interaction they have had with it.
Brand Mark Family
A brand mark family is the complete set of five to eleven logo variants built from one visual logic so the identity survives every surface from hero banners to sixteen pixel favicons without manual rescue.
Brand Mascot
A brand mascot is a character built to embody your brand personality, values, and promise so users form an emotional connection that sticks. When executed with discipline it becomes a working asset that drives notifications, content, retention, and recognition across every touchpoint.
Brand Motif
The repeating visual element that identifies the brand when the logo is absent. It is the pattern, shape language, illustration approach or photography treatment that makes every surface feel like the same brand at any scale.
Brand Neutral Scaffold
The generic unbranded UI structure AI tools generate from prompts that delivers correct layout hierarchy and component placement but zero trace of your specific colors typography or visual identity.
Brand Palette
The defined set of primary, secondary, and accent colors that represent a brand's visual identity across all touchpoints. More structured than a generic color palette.
Brand Personality
The specific human traits that translate your archetype into consistent daily behavior across copy, design, and interactions.
Brand Pillars
Brand pillars are the three sharp ideas your brand must prove across every asset and interaction. They sit in the strategy foundation and act as the filter that turns vague intentions into concrete verbal, visual, and systems decisions.
Brand Positioning
Brand positioning is the specific spot you claim in your target audience's mind that explains why they choose you over every alternative. It is the strategic foundation built from customer insight, competitive gaps, and your actual strengths.
Brand Promise
The core value or experience a brand commits to delivering to its customers, forming the fundamental expectation for every interaction.
Brand Refresh
A brand refresh sharpens the visual and verbal surface of an identity without touching the underlying strategy, positioning, or meaning.
Brand Reputation
The aggregated trend of brand image across many people over many years. It compounds every promise kept or broken.
Brand Rollout
Brand rollout is the sequenced execution that deploys a new identity across every internal operation product surface and customer touchpoint after the strategy and design work finishes.
Brand Strategy
The one-page foundation that defines who the brand is for, what it stands for, how it differs from alternatives, and what it must never be.
Brand System
The interconnected set of visual and verbal rules that work together to produce a consistent brand experience across every context.
Brand Systems Lead
The designer who builds living brand infrastructure so both humans and AI agents produce consistent work at massive scale without constant founder intervention.
Brand Teardown
The disciplined reverse engineering of a marketing site's five layers to name every framework, font, motion primitive, visual constraint, and performance choice that makes it read as current.
Brand Touchpoint
Any moment where a person interacts with or encounters a brand, from a website visit to a packaging unboxing to a customer service call.
Brand Typography
Brand typography is the strategic choice of typefaces that position a company, build trust, and drive conversions before customers read one word of copy.
Brand Voice
How a brand sounds in writing and speech. The personality, tone, and word choices that make it recognizable even without visuals.
Branded House
One master brand owns every product, extension, and touchpoint. Descriptors handle differentiation while a single voice, visual system, and design grammar carry the entire portfolio with zero sub-brands.
Branded Spinner
A branded spinner is the custom micro animation that layers product personality onto structured loading states using signature gradients, pulses, and timing curves pulled from your design system.
Breadcrumbs
Breadcrumbs are secondary orientation links that display the exact path through a site's hierarchy from root to current page. They collapse deep structures into one scannable line so users who land via search instantly understand where they sit and how to climb back out.
Brightness Test
The brightness test drops your built dark interface to 30 percent screen brightness in a pitch black room so borders disappear, text halos, accents vibrate, and elevation collapses exactly as they do for real users at night.
Brutalist Minimal
Brutalist minimal is the hero pattern that weaponizes the total absence of marketing polish. It delivers plain text descriptions and functional forms on an unadorned background to signal raw competence and filter for users who distrust conventional sales tactics.
Brutalist Web Design
Brutalist web design exposes raw structure, massive type, and visible grids to signal confidence without decorative polish.
Budget Kill Number
The exact dollar figure written into the design brief that ends all work the moment it is reached. It sits 15 to 25 percent above the sold price and is signed before the first pixel moves.
Built-in Social Proof
Built-in social proof places real user validation directly inside the product surface at the exact moment of trust friction instead of hiding it on a marketing page nobody reads.
Bulk Operations
Bulk operations let experts select dozens or hundreds of items and update them together instead of repeating clicks one by one. Linear's shift-click then B, Notion's database multi-select, and Gmail's keyboard shortcuts show how visible selection, keyboard actions, and real undo compound into a system experts cannot live without.
C
Cache Read
A billing event where an AI agent retrieves previously stored context from its cache, typically at a significantly reduced token rate.
Cache Write
A cache write is a billing event where an AI agent stores new or refreshed context into its temporary memory, incurring a higher cost than merely reading existing cached information. It's the premium you pay for giving your agent a persistent working memory.
Calibration Loader
A calibration loader is a branded system boot sequence that masks heavy asset streaming in interactive experiences while building anticipation and training users for the interaction ahead.
Camera Orbit
Camera orbit moves a virtual camera along a curved path through space around a completely frozen scene. The resulting frames deliver real parallax where foreground elements shift dramatically against barely moving backgrounds.
Canonical MCP Server
Canonical MCP servers are the nine official implementations from Figma, Linear, Notion, GitHub, Slack, Stripe, Postgres, Filesystem, and Browserbase that expose real resources, tools, and prompts through the Model Context Protocol so any 2026 AI client can read live data and trigger actions without custom glue code.
Capability Bound
The upfront map of exactly what an AI can and cannot do shown on the first screen so users calibrate expectations and build a working mental model before they type their first prompt.
Capacity Based Pricing
Capacity based pricing sets your productized design subscription fee by first calculating the exact number of clients you can serve without missing turnarounds then dividing your revenue target across that limit. It forces you to treat your own bandwidth and systems as the non negotiable constraint instead of copying competitors or chasing client budgets.
Capture Stack
The layered toolkit of capture tools, automation scripts, annotation layers, and organized storage that turns real product interfaces into a living, searchable, AI-ready design database.
Card Wall
A card wall is a uniform grid of identical cells that present parallel content at equal visual weight. Most 2026 implementations labeled as bento grids are actually card walls because they retrofit content into same-size containers instead of earning size through content density.
Case Study Reflection
The final paragraph in a case study where you admit what you would do differently next time, written with tactical specifics that show seniority and willingness to learn.
Categorical Color
Categorical color assigns a unique hue to each discrete category with no order or magnitude so the eye can separate unrelated groups at a glance. One of three strict color roles in dataviz, it demands no more than six distinct hues or the encoding collapses.
Category Color Inversion
Category color inversion is the deliberate rejection of a category's dominant palette to signal a different promise and attract the buyer the rest of the category ignores. It turns biological color expectations against the market itself so the brand registers before a single word of copy lands.
Category Contrast
Category contrast is the strategy of mapping every visual, verbal, and tonal convention in your product category then deliberately breaking the loudest ones with total commitment across every surface.
Change Order
A formal contract amendment that adds specific deliverables, adjusts timelines, and increases fees when the client requests work beyond the original signed scope.
Chat Shaped Hammer
The chat shaped hammer describes the tendency for teams to treat every AI problem as a nail solvable by bolting on a conversational interface even when direct manipulation or structured output would clearly be faster and more appropriate.
Chatbot
A turn-based conversational interface that responds to one prompt at a time with plain text and then waits for further instructions. It has no goals, no tool access, and no ability to plan or execute multi-step work on its own.
Checkmark Sea
The dense grid of green checkmarks, faint dashes, and random Xs that swallows pricing pages when teams cannot commit to clear opinions about their buyers or their value.
Chromatic Adaptation
Chromatic adaptation is the eye's rapid adjustment to a dominant color in your field of view. Within 500 milliseconds neutrals can appear cooler or warmer based on what came before shifting how every subsequent color reads.
Chrome Bloat
Chrome bloat is the hidden tax teams pay when they remove a sidebar only to rebuild equivalent weight in top bars, right panels, and floating controls. The interface feels busier, the content area shrinks, and users pay the price in cognitive load.
Circular Grid
A circular grid is a construction system of overlapping circles sized by golden ratio silver ratio or simple multiples that locks every curve intersection and proportion in arc-driven logos.
Citation Surface
A citation surface wires every AI claim to its exact source so users can verify output instead of praying the model did not hallucinate.
Classifier Layer
Anthropic's runtime safety system that scans every Fable 5 prompt and output for policy violations then silently downgrades the backend to Opus 4.8 on a refusal stop reason while returning HTTP 200.
Claude Code
Anthropic's agent-mode command-line tool that reads your entire codebase, edits files, runs tests, and opens pull requests from a terminal prompt.
Claude Design
Anthropic's AI tool that generates clickable HTML/JS prototypes from natural language prompts, leveraging ingested design systems for rapid, on-brand first drafts.
Claude Skill
A Claude Skill is a folder containing a SKILL.md file with YAML frontmatter and instructions that Claude loads automatically when a request matches its trigger description.
Claude Skills
Versioned packages of prompts, rules, reference examples, and evaluation logic that turn Claude into a repeatable design specialist. Top teams in 2025 used them to automate brand audits, enforce component systems, critique interfaces, and generate production code without losing craft.
CLAUDE.md
CLAUDE.md is the markdown file you place in your repo root that acts as a permanent design brief for Claude Code, encoding your brand voice, component rules, token paths, and quality standards so the agent behaves like a teammate who already gets it.
Cleanup Tax
The Cleanup Tax is the hidden cost of making AI-generated code production-ready, transforming visually correct but contextually naive output into something shippable and aligned with your design system. It represents the engineering and design effort required to bridge the gap between AI's raw output and your product's quality standards.
Client-Side Secrets
API keys, database credentials, and other sensitive values embedded in frontend JavaScript during rapid AI generation, visible to anyone who opens browser developer tools.
CMS
A Content Management System is the layer that lets non-developers create, edit, and publish structured content without touching code. It splits the people who write from the people who build, so a marketer can ship a blog post while the engineer is asleep.
Code Connect
Code Connect maps Figma components directly to their code twins so AI agents import real components with correct tokens instead of hallucinating new ones.
Code Gen UI
Interface created when a model writes fresh code on demand to produce layouts components or entire experiences instead of choosing from a fixed vocabulary.
Code Literacy
Code literacy is a designer's ability to read, understand, and edit components in a live codebase. It replaces Figma frames with direct work on the shipped product using tools like Cursor, v0, and real design tokens.
Code Reading
Code reading is the skill of scanning TSX files like Figma components by focusing on props variants states and styling to spot mismatches before they ship. It lets designers review frontend PRs with confidence without ever writing production code.
Code-first Design
Code-first design treats the live codebase as the primary surface for product design instead of static Figma frames. Designers use v0, Cursor, Lovable, and real component libraries to build, iterate, and ship with perfect fidelity because the design and the implementation are the same artifact.
Coined Name
A coined name is a word built from nothing or fused so tightly the original pieces vanish, delivering total trademark ownership at the cost of teaching the world its meaning from absolute zero.
Cold Start
Cold start is the first real interaction a user has with an AI product. It drops them into their actual work immediately so the mental model forms around real tasks instead of demos.
Cold Traffic
Cold traffic is any first-time visitor with zero brand familiarity who lands from paid ads, unbranded search, or random social links. They owe you nothing and will bounce in seconds if your design makes them work to understand the offer.
Collaboration Case
A collaboration case is the portfolio project that shows you can ship good work alongside other people instead of in spite of them. It details the trade offs, compromises, and real team dynamics that every complex project contains.
Color Accessibility
Ensuring color choices meet minimum contrast standards so content is readable by users with color vision deficiencies. WCAG AA requires a 4.5:1 contrast ratio for normal text.
Color Build Order
The specific sequence for constructing a working palette so each layer constrains the next. Neutrals first, anchor second, semantic third, state fourth, dark last. Running it backwards creates most palette pain.
Color Constraint
A color constraint is a crisp written rule that dictates how a brand color must behave across every use case, replacing vague hex codes with specific guardrails like Atlassian's bold without being loud.
Color Harmony
The pleasing arrangement of colors based on their relationships on the color wheel. Complementary, analogous, triadic, and split-complementary are the most common harmony types.
Color Jobs
The five distinct functions color performs for a brand at once: shaping perception, building trust, creating urgency, signaling price, and driving recall.
Color Palette
The defined set of colors a brand uses across all materials, typically including primary, secondary, accent, and neutral colors.
Color Psychology
The study of how colors influence human perception, emotion, and behavior. In design, it is the practice of choosing colors based on the response they trigger, not just how they look.
Color Rationale
The documented paragraph that locks in why your exact brand color exists, which emotional and strategic jobs it performs, which audience insight it reflects, and which constraints it must never violate.
Color Selection Framework
A six-question decision tree that forces strategic color choices based on category conventions, emotional jobs, audience realities, technical limits, written constraints, and competitive positioning instead of picking whatever looks cool on a mood board.
Color System
A color system is a structured framework of roles, rules, and token mappings that dictate exactly how colors behave across states, surfaces, themes, and components rather than a loose collection of swatches.
Color Tokens
Named color variables tied to roles in a design system rather than raw hex values, so the same role can resolve to different colors in different themes.
Color Wheel
A circular diagram of hues organized by their relationships, used by designers to reason about which colors pair, clash, or complement based on their position on the circle.
Column Grid
The column grid splits a canvas into vertical strips of equal width separated by fixed gutters so every text block image and UI component snaps to the same invisible rhythm. It is the default layout engine for almost every marketing site app shell and content page shipped since Müller-Brockmann wrote the rules in 1961.
Combination Mark
A combination mark locks a wordmark or lettermark to a symbol in a ruled system so the brand can run the full lockup on big surfaces and the symbol alone on tiny ones without losing equity.
Command Bar
A command bar is the keyboard-driven parser that replaced the sidebar as primary navigation in modern SaaS. Hit cmd-K and it surfaces ranked actions, objects, and context-aware suggestions instead of forcing users to scan a static list of links.
Command Palette
A command palette is the single keystroke surface that combines search, navigation, and actions so experts can execute anything in the product without touching a menu or button.
Common Fate
Common fate groups elements that move together. The brain registers coordinated motion as one unit even when shapes colors and content differ completely.
Competitor Pass
The competitor pass is the first research step where you pull every competing and adjacent logo into one board then group them by color, mark style, era, and craft so you can see exactly what the category already owns and deliberately reject it.
Complementary
A color scheme built from two hues positioned directly opposite each other on the color wheel, generating maximum visual contrast and energy.
Component Composition
Component composition builds complex UI from small single responsibility pieces that combine like building blocks instead of creating monolithic components stuffed with configuration props.
Component Layer
The component layer is the top tier in the three-layer Figma token architecture. It holds specific tokens for component families that alias only to semantic tokens, creating an override point for product-specific decisions without polluting global rules.
Component Library
A collection of reusable UI elements (buttons, inputs, cards, modals) built from design tokens and documented with usage guidelines. One layer of a design system, not the whole thing.
Component Swap Pass
The mandatory process where designers replace every generic element from an AI scaffold like Figma Weave with real instances from their component library and design tokens.
Component Token
A component token is the most specific tier in a design token system. It overrides semantic defaults only for a particular UI element when that element needs a deliberate exception from the shared rules.
Component Tokens
The top layer of your token system that binds semantic roles like text-h3 to specific component slots such as card-title-size. This gives controlled overrides for real UI patterns without spawning anonymous pixel values that wreck consistency.
Component Variant
A component variant is a typed prop that switches one React component between distinct visual states defined in your design system.
Computer Use
Computer use is the AI capability that lets a model view screenshots, control a mouse and keyboard, and operate any software exactly as a human would.
Concept Direction
A one-sentence strategic North Star that defines exactly what the logo must communicate and the non-negotiable feeling it must carry before any sketch is drawn.
Conditional Rendering
Conditional rendering is the React pattern that swaps, shows, or hides UI elements using JavaScript logic inside JSX. The three shapes, ternary, short-circuit, and early return, each represent a visual state designers must own.
Confident Brutalism
Confident brutalism strips away every decorative safety net after the underlying design system has been hardened into something unbreakable. It uses massive type, exposed structure, and deliberate rawness to signal a brand that respects its audience enough to skip the gloss.
Confirmation Gate
The engineered pause that stops an AI agent before any irreversible or high-stakes action so it must present its exact intent in plain language and wait for explicit user approval, edit, or cancellation.
Content Fit
Content fit is the discipline of letting the weight and shape of a single idea dictate its bento cell size instead of forcing content into decorative boxes.
Content Reel
Content Reel is a Figma plugin that instantly fills designs with realistic names, avatars, phone numbers, addresses, and copy blocks instead of the same tired John Doe and lorem ipsum placeholders.
Context Discipline
The strategic management of an AI agent's entire working context, including system prompts, tool definitions, files, and conversation history, to minimize token costs. It focuses on optimizing the input the model processes, which accounts for the vast majority of AI agent billing.
Context Drag
Context drag is the increasing computational load, latency, and quality loss that happens when an AI model must reprocess a growing mountain of previous conversation on every new turn.
Context Drift
Context drift is the slow degradation of an AI coding agent's adherence to your design system and constraints as the session grows longer and the context window fills with new information.
Context Handling
The ability of an AI code tool to retrieve and use relevant code from across a repository rather than relying on raw context window size.
Context Principles System Proof
The rigid four-part presentation order that converts brand identity meetings from taste wars into confirmations of business decisions already made in the brief.
Context Rot
Context rot occurs when your opening prompt loses influence inside a long AI chat and the model begins optimizing for recent messages instead of your original role constraints and references.
Context Stability
Context stability is a model's ability to maintain consistent application of every design rule, brand guideline, component name, and voice constraint you set at the start of a session all the way to the final output instead of gradually forgetting or mutating them.
Context Threshold
The percentage of context window usage at which AI output quality begins to noticeably degrade, typically around 50-70% depending on the model and task complexity.
Context Window
The total amount of text, code, and conversation history an AI model can hold in active memory during a single session. Measured in tokens, not words.
Contextual Panel
A contextual panel is a right-side or overlay surface that shows controls and metadata scoped to whatever object you currently have selected instead of a global menu that never changes.
Contextual Settings
Contextual settings attach directly to the page, project, layer or object they control instead of hiding in a central registry. They kill distance between decision and action so users tweak behavior without leaving their flow.
Continuity
The eye follows the smoothest path through any layout. Alignments, edges, and curves become invisible rails that guide attention in one fluid motion so users scan instead of decode.
Contrast Ratio
The measured difference in luminance between two colors, used to ensure text and interactive elements are readable for all users.
Contribution Model
A contribution model is the explicit process, templates, review standards, and ownership rules that let product teams outside the core design systems group propose, build, and ship additions to the system.
Conversational Interface
Conversational interface is the chat based UI pattern that became the default for AI products after ChatGPT but forces every task into a single threaded dialogue even when that shape hurts speed, trust, and parallel work.
Conversational Queries
A post-dashboard pattern where the user types a question in plain language and receives a sentence-first answer with charts only when they clarify what words cannot.
Copilot
Copilot is inline AI assistance that lives inside tools like Figma, VS Code, or Notion and suggests your next move in real time while you stay in control.
Copy QA
Copy QA is the systematic audit of brand voice, tone consistency, and microcopy quality at scale using models tuned for taste instead of raw reasoning power. In 2026 it defaults to Claude 4.7 Sonnet paired with structured rubrics and custom skill packs.
Copy-Pasteable Artifacts
Copy-pasteable artifacts are AI outputs engineered as self-contained units that users can copy once and paste anywhere with full context, branding, source links, and remix paths intact.
Core Desire
The fundamental human motivation each archetype fulfills such as achievement for the Hero or transformation for the Magician.
Core Surface
Core surface puts the AI model at the center of the product as the primary interface users hit first. Remove the model and the product becomes a hollow shell instead of a working dashboard with one less button.
Core Web Vitals
Google's three measurable user-experience metrics for loading, interactivity, and visual stability that act as both a search ranking input and a design quality signal.
Covered Model
Anthropic's compliance category for frontier models like Claude Fable 5 that enforce a non-negotiable 30-day data retention policy on all inputs and outputs with no zero-data-retention option.
Creative Restraint
The discipline of removing elements, fonts, colors, and details until only what serves the core message remains.
Critique Facilitator
The critique facilitator owns time, focus, and capture during design reviews. They enforce frameworks and rules so the team delivers specific observations instead of opinions and the designer leaves with a clear path forward.
Critique Framework
A critique framework is a scripted structure for design reviews that forces feedback through observation interpretation and suggestion layers so teams surface blind spots and ship sharper work instead of trading vague opinions.
CTA
Call to Action. A design element, usually a button or link, that prompts the user to take a specific action like signing up, buying, or downloading.
CTA Ladder
A CTA ladder layers calls to action by buyer readiness so every visitor gets the right next step without noise or distraction from the primary goal.
CTA Microcopy
Short reassurance text under a button that kills the visitor's final objection before they click. It uses specific product facts to reduce risk instead of hype or urgency.
Cumulative Layout Shift
Cumulative Layout Shift (CLS) measures unexpected visual movement on a page after rendering begins. Anything over 0.1 makes your brand feel unstable and cheap while scores under 0.05 read as engineered precision.
Curated AI Work
AI-generated output that has undergone significant human judgment, transformation, and refinement to meet specific brand, audience, and quality standards.
Curation Diet
A curation diet is the daily practice of selecting one narrowly scoped design piece, annotating exactly what earns its place and what must die, then reviewing the library weekly to turn passive exposure into high-resolution pattern recognition.
Curation, Not Generation
The shift in a designer's primary role from creating every pixel and line of code to critically evaluating, refining, and guiding AI-generated design outputs. It emphasizes human judgment and taste over mechanical production.
Cursor
Cursor is the VS Code fork with a native AI layer that reads your entire codebase as context and executes natural language edits, file writes, and terminal commands.
Custom Axis
A custom axis is a four-letter uppercase tag defined by the type designer that opens a new dimension of continuous variation in a variable font beyond the five registered axes.
Custom Typeface
A custom typeface is a bespoke or semi-bespoke font family built specifically for one brand to create visual ownership, solve technical problems, and escape per-seat licensing costs.
D
Dark Canvas
Dark canvas is Linear's permanent brand decision to ship every surface on a near-black background with zero light toggle. It acts as both identity and filter, forcing typography, motion, and color decisions while telling terminal-dwelling engineers this tool belongs in their stack and everyone else can keep scrolling.
Dark Mode
Dark mode is a complete second theme with its own surface colors, contrast rules, elevation logic, and typography adjustments designed from scratch against the same components as the light theme.
Dark Mode First
Dark Mode First means building your primary UI in dark theme so every color and contrast decision earns its place under the toughest conditions before creating the light inversion.
Data Graveyard
A dashboard stuffed with widgets, metrics, and charts that nobody uses because it fails to tell the user what to look at first or what to do next.
Data Parity
Data parity ensures staging holds records with the same length, mess, volume, and edge cases as production so designers catch real breaks instead of signing off on fake perfection.
Data Visualization
Data visualization is the visual encoding of numbers into shapes that the eye can compare faster than the brain can read text. It turns raw metrics into charts, graphs, and dashboards that communicate insights at a glance instead of forcing users to parse tables of numbers.
Dead-End Settings
Dead-end settings are toggles that promise customization but deliver zero visible impact on the interface. They destroy user trust by making the settings page feel like a collection of broken promises and unfinished work.
Decay Timer
A visible countdown or event trigger attached to each memory card that tells users exactly when stored information will stop affecting agent behavior and gives them one-click power to change it.
Decision Log
A decision log is a six-section case study format that records every major choice a designer made from problem to shipped outcome instead of presenting a gallery of final screens.
Decision Stack
Decision Stack is a late-stage critique framework that requires the designer to present a crisp decision required, the options already considered, their recommendation, and every realistic risk before reviewers open their mouths.
Declarative Headline
A declarative headline makes one unambiguous claim about what your product is and who it is for. It filters the right audience in the hero before trust or action ever enter the picture.
Dedicated Ownership
Dedicated ownership means assigning a specific cross functional team to maintain, evolve, and govern the design system instead of treating it as a part time side project.
Deletion Test
The deletion test strips every model call from a product to reveal whether the AI was load bearing or just decorative trim.
Dense Data Design
Dense Data Design packs complex information into clean interfaces without forcing users to choose between clarity and craft. It uses tight typography, purposeful whitespace, and restrained signaling so that fifty data points can feel calm instead of chaotic.
Depth Piece
The third project in a tight five project portfolio where you prove you own an advanced specialty that generalists cannot touch. It is shorter than the flagship but packed with token schemas, research protocols, code snippets, and architecture diagrams instead of polished hero shots.
Depth Z-Stacking
Depth z-stacking treats the z-axis as a structured layout grid with strict elevation tokens that dictate exact shadows, blur, specular highlights, refraction, and parallax for every level from surface to overlay.
Desaturation Rule
Drop accent chroma 15 to 25 percent and text plus borders even more when building dark mode palettes. The adjustment prevents haloing, bloom, and eye strain at night while keeping hue locked so the brand stays instantly recognizable.
Design Code Convergence
The 2026 closing of the gap where designers began shipping production code and engineers internalized visual and interaction craft, eliminating the handoff layer entirely.
Design cold start
The challenge of starting a design project from a blank canvas, often for non-designers or vague concepts, where the goal is rapid generation of a credible first draft rather than polished production.
Design Comparison
Design comparison means placing every viable candidate next to the brief, the prior version, and the best work in the category so the gaps scream. The editors eye trains on contrast, never on isolated review.
Design Constraints
The real world limits of a project, listed clearly so readers understand the box you had to design inside instead of pretending you had unlimited time, budget, and freedom.
Design Critique
Design critique is the senior taste and judgment layer that survived the AI collapse of the old senior PD role. It is the decisive human call on ambiguity after Claude ingests the brand brief, prior critiques, and design system then delivers its first pass.
Design Diary
A design diary is a public six-field record of every major design decision with dated entries that list real alternatives, binding constraints, unhedged reasoning, named owners, and post-ship outcomes.
Design Drift
Design drift is the gap that opens between your Figma comp and production code when files lack strict token discipline and complete component variants. It turns intentional design into unintentional approximation through hundreds of micro decisions no one owns.
Design Engineer
A designer who ships production code, owns the living design system as the single source of truth, and closes the loop from taste to live product at companies like Linear, Vercel, and Anthropic.
Design Engineering Toolchain
The design engineering toolchain is the integrated stack a single design engineer uses to take a feature from Figma Variables through tokens and components all the way to a shipped Vercel preview. It eliminates handoffs by making every layer connect without manual translation steps.
Design Evals
Design evals are the automated tests defined in a spec that confirm the built feature matches the written intent, moving quality enforcement from manual QA into the spec itself.
Design Fidelity
Design fidelity is the degree to which a live website matches the original Figma design in layout, typography, animation, and interaction details without translation compromises.
Design For Ten
Design For Ten is the craft of building digital tools for a known group of roughly ten specific people whose names and needs you know by heart. It prioritizes uncomfortable precision and deep context over scalable onboarding and generic appeal.
Design Generalist
A designer who spreads across UI, brand, research, and strategy without owning any system, shipping production code, or developing a sharp specialty. In 2026 this became the least defensible role in the industry, a polite euphemism for expensive junior.
Design Governance
The ownership structure, decision-making process, and contribution model that determines how a design system evolves. The most common reason design systems fail.
Design Grid
A structured framework of intersecting lines used in design to organize elements, ensure alignment, and create visual consistency and balance.
Design Handoff
The structured transfer of a finished design from designer to engineer (or to the client's internal team), including source files, tokens, specs, and the open questions the recipient needs answered before they can build.
Design Infra Team
A dedicated product team that owns the design system as a first class internal product with its own roadmap, releases, and metrics. They keep tokens, primitives, and documentation in sync so product teams can ship instead of fighting drift.
Design Pattern
A documented solution for a recurring design problem. Patterns tell you when to use a dropdown versus a radio group, how to handle empty states, and how forms should behave.
Design QA
Design QA uses AI agents wired to Figma via MCP to compare live sites or code against your actual Figma frames and output exact token-level difference reports instead of vague vibes.
Design Queue
A design queue is the enforced list of client requests in a productized design service where work enters through standardized briefs, processes under strict rules for scope and one active request per client, and ships on a fixed turnaround.
Design Rationale
Design rationale is the documented logic explaining why specific design decisions were made. It connects original briefs, research data, metrics, and rejected alternatives to final choices in Figma or code and lets Claude maintain alignment through MCP servers pulling live context from Notion, Linear, and GitHub.
Design Reduction
Design reduction is cutting an interface until it breaks then restoring only the smallest element that makes it work again. It is the first move of the editors eye and the only reliable way to turn infinite AI output into work that matters.
Design Reframing
Design reframing rewrites the problem statement itself before any prompts fire or pixels move. It catches briefs asking for landing pages when the company actually needs positioning or mental model shifts and sits as the highest leverage move of the editor's eye.
Design Restraint
Design restraint is the discipline of subtracting every visual element that fails to serve the core story until only the essential remains.
Design Retainer
A design retainer is a monthly contract where a client pays a fixed fee to reserve a block of the designer's hours. It sells time instead of output which caps your income at the exact number of hours you can personally deliver.
Design Review Doc
A dated six-field document that records one design decision with its alternatives, constraints, reasoning, owner, and measured outcome so critiques target judgment instead of pixels.
Design RFC
A Design RFC is a short dated document that locks in a design decision before work starts by listing the exact call, live alternatives, binding constraints, unhedged reasoning, owner, and later outcome.
Design Spec
A design spec is a concise markdown document that encodes user intent, behavior, edge cases, success metrics, and acceptance criteria so both humans and AI can build the right thing without ambiguity.
Design Surface
Design surface is the visible layer of any TSX file that directly controls variants, states, layout, spacing, and composition. It maps one-to-one with Figma components while leaving hooks, effects, and data fetching in engineering territory.
Design System
A design system is the living product of tokens, components, patterns, guidelines, and governance that stops teams from reinventing UI every sprint.
Design System Cosplay
When teams build the appearance of a mature design system without the staffing, scale, or ownership required to make it useful.
Design System Maintenance
Design system maintenance is the endless work of keeping tokens, components, and patterns consistent across repos as the product evolves, the exact work that Claude Code now compresses from weeks into minutes.
Design System Migration
Design system migration is the process of remapping every token, component, and pattern from an old design language to a new one across an entire product. Claude Opus 4.8 makes it one session instead of file-by-file torture by holding both systems in its 1M token window and producing the complete mapping without losing thread.
Design Taste
Design taste is the judgment that cuts through ambiguity after AI ate synthesis, polishing, specs, and handoffs in 2025. It is knowing which generated option actually ships value, respects attention, and compounds over time when every variant looks viable.
Design to Code
Design to Code feeds real Figma structure into AI agents like Claude Code through MCP so the output pulls your exact tokens, components, and auto layout values instead of guessing from screenshots.
Design Token
A named variable in a design system that stores a visual decision (color, spacing, type, radius, motion) and can resolve to different raw values per theme, platform, or context.
Design Tokens
The atomic design values (colors, spacing, typography, shadows, motion) stored as platform-agnostic variables that every component in a design system references.
Design-system read
A design-system read is an artificial intelligence's ability to ingest an existing design system's codified rules, tokens, and components, then apply them to generate new, on-brand design outputs. This ensures AI-generated drafts are consistent with a brand's established visual language and technical specifications.
Design-to-Code Gap
The chasm between a finished design artifact and its functional, production-ready code implementation, historically a major bottleneck in product development.
Designed Friction
Designed friction is the deliberate placement of pauses, confirmations, and extra steps at high-stakes moments to protect users from accidental actions and signal respect.
Designer-PM
A senior design IC who owns the written spec, named metric, ship date, user research, and launch write-up on top of the interface work. The role that ate the old PM IC seat in 2026.
Designing In Code
Designing in code means treating a live codebase as your primary canvas in 2026. Designers edit tokens and components directly with AI pair programmers like Claude Code and Cursor then ship deployed surfaces instead of static Figma frames.
Desktop-Only Scale
A desktop-only scale is a typography system failure where type rules, optimized for large screens, break down and create a poor user experience on smaller devices. It's a design oversight that neglects responsive behavior, leading to illegible text and inconsistent brand perception.
Destructive Regenerate
Destructive regenerate is the anti-pattern where an AI interface overwrites and deletes the prior streaming response when the user hits regenerate, erasing partial wins and forcing a full reset instead of preserving history for comparison or synthesis.
Detached One-Off
A detached one-off is a component instance ripped from its source definition in Figma or similar tools, creating hardcoded local edits that contradict the design system and become equally valid signals for any machine reading the file.
Dev Environment
Dev environment is the local sandbox where engineers build and test code on their laptops using fake data and temporary setups. Review here for big picture structure and behavior but never for final polish or client approval.
Didone
Didone is a high contrast serif category named after Didot and Bodoni. These faces use mechanical precision and violent stroke contrast to communicate exclusivity and craft at a glance.
Direct Labeling
Place labels on the data itself instead of using legends that force eye ping-pong between key and chart. It cuts cognitive load so the user reads the story in one uninterrupted pass.
Direct Manipulation
Direct manipulation is the AI interface pattern where the model watches what you are already doing in your editor or canvas and offers inline suggestions you accept with a single keystroke or click instead of routing everything through a chat thread.
Direct to Canvas
Direct to canvas opens products straight to the workspace, inbox, or prompt bar so the user creates before they notice the missing tutorial. The interface borrows from familiar patterns so the product itself does the teaching in 2026 AI tools.
Discovery Document
The 4 to 8 page document that captures unfiltered client truths about what the business actually does, who it serves, and where it is headed, then calls out the core tensions the brand identity must resolve. It gets signed off before anyone opens Figma.
Discovery Moment
The discovery moment is the exact instant a viewer decodes the hidden secondary meaning in a negative space logo and their brain snaps the two reads together. That click releases a dopamine hit that glues the brand deeper into memory than any direct symbol ever could.
Display Typeface
A display typeface is engineered for large sizes where high contrast, sharp details, and strong personality command attention and set brand tone.
Distribution by Design
Distribution by Design is the practice of engineering share loops directly into the core product surface so every AI output becomes a self-contained ad unit that travels on its own. Design teams own virality by composing screenshottable surfaces, demo-first flows, copy-pasteable artifacts, built-in social proof, and share-multiplier outputs instead of bolting on social features after the fact.
Diverging Color
Diverging color uses two contrasting hues separated by a meaningful neutral midpoint to encode values running from negative to positive or below to above a fixed target.
Do Not Panel
An annotated visual section in a brand book that shows realistic off-brand executions with red arrows, callouts, and blunt labels so teams instantly recognize and avoid the same mistakes.
Done Condition
The specific artifact, live URL, or shipped surface that proves a design project has reached completion. It replaces we will know it when we see it with a screenshotable deliverable everyone can agree on.
Dragging Movements
WCAG 2.2 criterion 2.5.7 requires every drag or swipe gesture to offer a click, tap, or keyboard alternative so motor-impaired users are not locked out.
Draw Animation
Draw animation traces a logo's strokes in sequence to show its construction. It succeeds only on marks with strong line geometry and looks like a mistake on solid wordmarks.
Dual User Audit
The dual user audit is a seven question checklist that verifies your interface works for both the human clicking around and the agent parsing your DOM at the same time. It forces stable names, structured states, visible diffs, and documented surfaces so agents can drive your product instead of guessing at it.
Dwell Time
Dwell time is the number of seconds a visitor actively engages with a specific section before scrolling away or leaving. It tracks real attention through mouse movement, scroll pauses, and hovers instead of passive tab time.
E
Earned Scale
The principle that type or elements only get to be large when the content justifies the real estate through importance, impact, or meaning.
Editor's Eye
The muscle of selection, refinement, and judgment that becomes the designer's entire job once AI makes generation free and infinite.
Editorial Restraint
Editorial restraint pairs dense literary copy and sophisticated serifs with purposeful craft details that trust the audience to read and think. It turns restraint into a signal of confidence instead of the fear driven emptiness that defines the 2026 sameness dialect.
Editorial Rhythm
The deliberate pacing across spreads or screens that makes a layout feel like music instead of a random sequence of loud elements.
Effort Dial
The /effort command in Claude Code that lets you set how deeply the model reasons before answering, from low for instant tasks to ultracode for multi-agent builds.
Effort Levels
Claude's five settings that trade token spend for deeper internal reasoning. Low effort is fast and cheap but shallow. Max effort spends ten times more to unlock the long horizon thinking Fable 5 is known for.
Eject Penalty
Eject penalty is the brutal tax in time, money, and momentum you pay when you outgrow a hosted AI app builder like Lovable and try to run the code anywhere else. The magic dies the moment the exported app collides with real infrastructure, real auth needs, or real scale.
Embedded Analytics
Embedded analytics places data directly inside the working surface where decisions happen so metrics become native properties of the object instead of a separate destination.
Emblem
An emblem seals the brand name and central symbol inside one rigid enclosed shape such as a circle, shield, or badge. The format projects instant heritage, authority, and permanence which is why it dominates cars, colleges, coffee roasters, and beer labels.
Empty State
The screen a product displays when it has no data or content to show. It serves as the activation surface that determines whether a new user returns for a second session.
Empty State Archetype
Empty state archetypes are the five distinct blank-screen scenarios every product must solve: first run, no results, no permission, no connection, and no AI output. Each demands its own copy, hierarchy, action, and posture or the product leaks users before they ever see the real value.
Empty State Audit
A seven question checklist that validates whether an empty state is ready to ship or still leaks users and retention.
Endorsed Brand
Sub-brands keep their own names, personalities, and visual systems while the parent supplies a visible trust signal through lockups and shared grammar.
Engineering Surface
Engineering surface is the layer of TSX code that handles state hooks, side effects, async data fetching, and server logic. Designers learn its four distinct shapes so they can skim past it without anxiety and keep every PR comment locked to the actual design decisions.
Environment Variable
Environment variables are named values injected at runtime so the same codebase can connect to different databases, APIs, and services across dev, staging, and production.
EPS
EPS is a legacy vector format from 1982 that wraps PostScript code so old print equipment can output crisp scalable logos. It survives today only because some vendors still run 20 year old RIP software and refuse to switch to PDF.
Error Recovery
Error recovery is the UI surface that activates when an agent step fails. It surfaces the exact failure point, preserves all prior state, and offers concrete buttons for retry, edit, manual takeover, backtrack, or handoff so the task moves forward instead of dying.
Error-Empty
The empty state caused by a failure to load data, dropped connection, or misconfigured integration. It must explain the problem in plain English and offer a clear recovery path to maintain trust.
Eval Loop
The eval loop is the generate-eval-tune cycle that catches brand drift in real time so the editor fixes the system instead of fixing ten thousand individual assets.
Eval Pyramid
The four-layer filter that turns thousands of AI design candidates into the few worth shipping: deterministic lint at the base, visual diff and regression, LLM-as-judge running a structured rubric, and human taste reserved for the top with conversion data closing the loop.
Eval Stack
The four-layer system of cheap deterministic checks, visual regression, LLM-as-judge scoring, and human taste review that filters AI-generated design candidates before anything ships.
Explicit State
Explicit State turns every user interaction into a named variant on a single component so machines can read default, hover, focus, disabled, and error without guessing or measuring pixels.
Exploration Round
The presentation of three to five distinct creative directions, each answering the strategy brief in a different way, shown as printed concept pages not polished comps.
Expressive Display
Expressive display typefaces deliver maximum personality through extreme forms, wild details, and loud proportions built exclusively for hero sizes and short term campaigns.
External Memory
External memory is the durable knowledge you store in files, docs, and notes outside any AI chat so every session starts clean and performs at full strength.
Eye of Sauron
The Eye of Sauron is a real-time unprioritized feed that dumps every event, log line, webhook, and system ping onto the screen with zero hierarchy. It creates the illusion of complete visibility while turning every signal into noise and every decision into guesswork.
F
F-Pattern
A common eye-tracking pattern where users scan content-heavy pages in an F-shape: across the top, down the left side, then across shorter horizontal scans.
Fast Mode
Fast mode runs full Opus 4.8 quality inside Claude Code at usable speeds without dropping to a smaller model. It kills the old tradeoff where designers chose between smart answers and fast answers.
Favicon
The small icon displayed in browser tabs, bookmarks, and mobile home screens. Typically 16x16 or 32x32 pixels, it is the ultimate scalability test for any logo.
Feature Flag
A feature flag is a runtime control that determines which users see a new design in production. It lets teams deploy code to everyone while exposing the experience only to selected accounts.
Feature Row
The three-column equal-weight layout that dominated SaaS marketing pages from 2018 to 2023 by presenting features in identical boxes with icons, headings, and short copy.
Few-Shot Example
A few-shot example is a set of three to five real before-and-after pairs pulled from past team work and baked directly into a prompt so the model copies proven taste instead of guessing at vague rules.
Figma AI
Figma AI is Figma's integrated suite of artificial intelligence features, designed to accelerate professional design work directly within the Figma canvas, focusing on iteration, refinement, and production readiness.
Figma Dev Mode
Figma Dev Mode is the native inspector that turns design files into engineer-ready specs with live measurements, token-aware values, generated code, and direct links to code components. It killed most third-party handoff tools after the 2024 updates and became mandatory for any team maintaining a real design system.
Figma Make
Figma Make is the 2026 AI feature that converts structured Figma frames into production-ready React components using shadcn, Tailwind, and your exact design tokens.
Figma MCP
Figma MCP is the official local server Figma shipped in 2025 that feeds your real file structure, components, and design tokens directly to AI agents like Claude Code through the Model Context Protocol.
Figma MCP Integration
Figma MCP Integration is a Model Context Protocol server that exposes your live Figma components, semantic tokens, and design rules to AI coding tools like Cursor, Claude Code, and Windsurf so generated code matches your system on the first try.
Figma Round-Trip
The full workflow of generating a UI layout with AI then importing it into Figma for token remapping, structural refinement and eventual developer handoff.
Figma Template Effect
The homogenization that occurs when brands build from the same popular Figma community files, component libraries and defaults until every homepage shares the same DNA.
Figma to Code
Figma to Code connects your Figma designs to Claude Code via the Model Context Protocol so the AI reads your real design data and generates components that match your tokens and library instead of guessing from screenshots.
Figma Variables
Figma Variables replaced styles as the single source of truth for design tokens in 2025. They hold named, mode-aware values for colors, spacing, radii, and typography that feed straight into Tokens Studio and your production codebase.
Figma Weave
Figma's native AI tool that converts plain-text prompts into editable layout scaffolds dropped directly onto your canvas as real layers.
Figmotion
Figmotion is a Figma plugin that adds a full keyframe timeline for animating layers with precise control over position, scale, rotation, opacity, and paths plus one-click Lottie export.
Figure Ground
Figure ground is the perceptual principle that controls whether your eye reads a shape as the main subject or the supporting background. In logos it turns negative space into a second deliberate image that reinforces the brand truth without adding extra elements.
First Five Minutes
The decisive period where users test an AI product's first outputs for visible reasoning, citations, model limits, and reversibility to decide if it is a trustworthy tool or another hallucinating gimmick.
First Prompt
The first prompt is the initial real input a user gives an AI product that delivers immediate value and locks in their mental model of what the tool can and cannot do.
First Read
The element that grabs the eye in the first second and owns the highest rank in visual hierarchy. It is the headline value prop or CTA you engineer with size contrast and space so visitors absorb the core message before they bounce.
First Thirty Seconds
The first thirty seconds is the cold open where an AI product must prove its core value to a new user without narration setup or handholding. It functions as the silent marketing video that drives shares because most discovery happens through muted screen recordings on timelines.
First-Run Empty
The empty state a new user sees immediately after signup before they have created any content or data. It functions as the product's actual first impression and primary activation surface.
Five-to-One Rule
The five-to-one rule states that one hour of sharp human direction should yield the equivalent of five hours of AI-generated output for review and judgment. It is the precise ratio that turns AI from a novelty into the leverage engine powering AI-native design teams in 2026.
Fixed Canvas
A design surface with immutable physical dimensions and real production costs, forcing decisions that digital screens often dodge.
Flagship Project
The flagship project is the lead case study on your portfolio homepage that demonstrates full ownership from research to launch and hooks hiring managers in their 90 second scan.
Floating Labels
Labels that begin as placeholder text inside an input field and animate upward on focus to sit above it, attempting to solve both space and context at once.
Fluid Typography
A responsive technique where font sizes scale smoothly between breakpoints using CSS clamp(), eliminating the layout jank from hard breakpoint changes.
Focal Point
The first element the eye is drawn to in a composition. Established through size, contrast, color, or isolation, a focal point anchors the entire visual hierarchy.
Focus Not Obscured
WCAG 2.2 AA criterion 2.4.11 that requires the keyboard focus indicator to stay at least partially visible in the viewport and never get swallowed by sticky headers, cookie banners, floating widgets, or overlapping modals.
Focus Ring
A focus ring is the outline or box-shadow that appears around interactive elements when they receive keyboard focus. It must hit a 3:1 contrast ratio against its background per WCAG 2.2 non-text rules or users tabbing through your interface become lost.
Focus State
The visual indication that an interactive element currently has keyboard focus, required by WCAG 2.2 and the only way keyboard and screen-reader users know where they are on a page.
Font Axis
A font axis defines a continuous range of styles inside a variable font so one file can render any value along that dimension instead of forcing you to load separate static masters.
Font Budget
Font budget is the hard cap on typography assets that keeps LCP fast and brand perception premium. Two weights max, aggressive subsetting to only needed characters, and loading strategies that never block render or shift layout.
Font Buffet
A common anti-pattern in typography systems where a design uses too many typefaces without a clear strategic rationale, leading to visual chaos and brand inconsistency.
Font Pairing
The strategic selection of two or more typefaces that work together in a design system. Good pairings create contrast in structure while maintaining harmony in proportion.
Font Preload
Font Preload is the link rel=preload as=font tag placed in the document head that forces the browser to fetch your Latin variable font subset with high priority before it finishes parsing CSS.
Font Subsetting
Font subsetting trims a font file to only the exact glyphs, language ranges, and variable axes your project uses. It turns a 580 KB Recursive variable monster into a focused 140 KB file that still interpolates across every weight, slant, and custom axis you actually declare in CSS.
Font Variation Settings
Font variation settings is the CSS property that controls specific values along each axis of a variable font at runtime. Instead of hoping font-weight lands on the right stop you set exact numbers for wght wdth opsz or any custom axis the designer exposed.
font-display
font-display is a CSS @font-face descriptor that controls what the browser shows while a web font is still downloading. Its five values, auto, block, swap, fallback, and optional, decide the tradeoff between invisible text, a fallback flash, and layout shift. The common recipe is swap paired with a metric-matched fallback, or optional when you cannot afford any shift at all.
Forced Reps
Forced reps are the weekly practice of shipping one complete design piece without AI, submitting it to a fifteen-minute critique from a sharper eye, then rewriting every note in your own words as principles.
Forcing Function
A forcing function is a deliberate UX mechanism that blocks a high-stakes action until the user actively proves intent through behavior like typing a exact resource name or reviewing every line of an AI diff. It converts muscle memory into conscious choice at moments where accidents would be irreversible.
Form Autocomplete
Form autocomplete is the autocomplete attribute that tells the browser and password managers exactly which slice of saved user data belongs in each field turning manual entry into instant population.
Form Gate
Form Gate is any multi-field input sequence that blocks the first AI prompt so the product can collect role, company size, use cases, and other context first. It treats onboarding like a CRM data entry task instead of the mental model problem it actually is.
Form Microcopy
Form microcopy is the short targeted text that instructs reassures and directs users within web and mobile forms. It turns potential points of friction into smooth interactions by answering questions the moment they arise.
Format Decision
The format decision is picking which of the seven structural containers (wordmark, lettermark, pictorial mark, abstract mark, mascot, combination mark, emblem) the brand will live in before any sketching starts. This choice dictates how the mark performs at 16 pixels, in one color, on embroidery, and across every unsupervised touchpoint.
Founding Designer
The first design hire at an AI-native startup who owns brand, product, marketing site, and design system in one seat while reporting directly to the founder. Equity-heavy, taste-driven, and built for unbounded upside when the fit is right.
Four Question Framework
A sequential set of four questions that replaces quizzes to select a brand archetype by defining promise, feeling, enemy, and voice.
Four-Layer Figma File
A four-layer Figma file organizes your designs into tokens, components, patterns, and pages so developers and AI coding agents can implement them without guesswork or drift.
Four-Tier Wall
The four-tier wall is what happens when a team cannot commit to a customer segment so they build four pricing columns of equal height with overlapping features. Hobby, Pro, Team, and Enterprise sit there like identical soldiers while the visitor hunts for a difference that does not exist.
Frame Ghosting
Frame ghosting is the translucent double image that appears when you blend two frames with mismatched parallax instead of snapping to the single nearest frame at full opacity.
Friction Audit
A friction audit is a four-question review that forces every high-stakes flow to justify its pauses so the product protects users from expensive mistakes instead of rushing them into regret.
Friction Engineering
Friction Engineering is the ruthless deletion of every unnecessary field, click, delay, and visual element that taxes conversion on landing pages. In 2026 it means one-field forms or SSO, progressive disclosure via bento grids, and sub-800ms load times because each extra input or millisecond directly kills signups.
Frontier Model
The handful of flagship AI systems that set new standards for reasoning, context handling, tool use and reliability in any given year.
Full Context Synthesis
Full context synthesis feeds every relevant design artifact into one AI session so the model reasons from the complete picture instead of a sampled slice that misses critical patterns and creates drift.
Full Lockup
The full lockup is the formal flagship version of a brand logo containing the mark, complete wordmark, and tagline. It exists exclusively for large-scale applications starting at 200 pixels where every element has room to breathe and assert authority.
Full-Bleed Canvas
A full-bleed canvas strips persistent navigation rails from the screen and lets the primary artifact run edge to edge. The work itself orients the user while a command bar and gestures handle jumps.
Functional Identity
Functional identity embeds brand assets directly into product workflows so every color, illustration, and word solves a user problem while reinforcing recognition.
G
Gamut Clamping
Gamut clamping is what browsers do when an OKLCH color exceeds the sRGB boundary. The engine projects the color to the nearest displayable value inside the gamut which usually rotates the hue even though you never changed the H coordinate.
Generative Summaries
Generative summaries synthesize raw data, events, and context into specific narrative paragraphs that lead with insight and recommended actions. They replace chart grids with server-side synthesis that reads like your sharpest teammate already digested the numbers.
Generative Surface
Generative surfaces are on-demand UI panels that AI assembles from prompts or selections instead of static pages wired in Figma months earlier. Tools like Cursor and Granola use them to replace fixed taxonomies with dynamic composition that only exists the moment you need it.
Generative UI
Generative UI is the practice of prompting AI models to output complete interface layouts, components, and working code from natural language descriptions instead of manually drawing them in design tools.
Generative Vocabulary
The complete system of primitives, intent slots, fallback states, recoverability affordances, and citation rules that lets an LLM compose brand-consistent interfaces at runtime. It turns design from drawing fixed screens into defining the exact grammar the model must speak.
Geometric Foundation
The practice of building logos, especially minimalist ones, using fundamental geometric shapes like circles, squares, and triangles for universal understanding and structural integrity.
Geometric Grain
The faint dot grid, hairline pattern or noise texture added to dark surfaces to create depth and sophistication without introducing new colors. It started as Vercel's signature move in 2024 and became the fifth tell of the SaaS monoculture by 2025.
Geometric Grotesque
Geometric grotesque is a sans-serif typeface built from pure geometric forms with even stroke weights, closed apertures, and neutral proportions that deliver precision and longevity for tech and product brands.
Geometric Logo
A logo built primarily from circles, squares, triangles, or other mathematical shapes. Geometric foundations give marks universal readability and structural integrity.
Geometric Sans
Geometric sans is a sans-serif category built from perfect circles, squares, and straight lines with even stroke weights and zero calligraphic influence.
Gestalt Closure
Gestalt closure is the brain's instinct to fill missing visual data and perceive incomplete shapes as whole forms. In logos it lets a few precise marks plus strategic negative space imply an entire recognizable figure.
Gestalt Principles
The eight perception rules the brain runs to group pixels into meaning before conscious thought arrives. Interfaces at Linear Stripe Figma and Notion obey them. Everything else makes users do the work manually.
Glassmorphism
Glassmorphism is the frosted translucent UI style that dominated 2021-2023 interfaces with blurred backgrounds, low-opacity panels, and bright borders but now serves as a dated marker for template-driven AI startup sites.
Goal Rationale Feedback
Goal Rationale Feedback is a critique framework where the designer states a crisp goal in two sentences max, explains their rationale in under three minutes, then receives all feedback tied directly to that goal or parked for later.
Golden Ratio
Golden ratio is the 1.618 scale multiplier that generates dramatic type sizes from a single base, turning 16px body text into 26px, 42px, 68px, and 110px headings that command immediate attention.
Golden Ratio in Logos
Golden ratio in logos uses the 1.618 proportion as a sanity check on relationships between shapes rather than a literal generator that dictates every single measurement.
Graceful Refusal
Graceful refusal is an AI product's clear admission that it lacks the data, context, or capability to answer reliably instead of hallucinating a confident but wrong response. It turns potential trust breakers into trust deposits by showing the system knows its own edges.
Gradient Mesh
The blurred coral-to-violet-to-cyan radial gradient layered behind hero text on dark SaaS homepages. Once a premium signal from Stripe and Linear, it became the default that every AI tool spits out and every founder copies.
Gradient Trap
A gradient trap is a logo that relies on a custom color blend to fake form, depth, and meaning. Strip the gradient and the mark collapses into a generic or broken shape that fails the single-color test and most other audit questions.
Grotesque
Grotesque is the original industrial sans-serif, the 19th-century family that dropped the serifs, held the stroke nearly uniform, and closed up the apertures for a flat, neutral tone. Akzidenz-Grotesk (1898) set the template, and the Helvetica and Univers lineage refined it into the default voice of modern branding.
H
Halation
Halation is the glowing blur around pure white text on pure black backgrounds that fatigues the eye and destroys readability in dark environments. It happens because extreme contrast overwhelms the visual system when pupils are dilated.
Hallucinated Confidence
The anti-pattern where an AI delivers uncertain or fabricated outputs in the same authoritative tone it uses for verified facts, with zero caveats, sources, or visible reasoning.
Hallucinated UI
Hallucinated UI is when a generative model invents buttons, tabs, charts or entire controls that render perfectly but have no backend handlers, real data or actual function.
Hamburger Menu
The hamburger menu is three stacked lines that conceal navigation behind one tap on mobile. It saves screen space yet tanks discoverability and makes every hidden item feel secondary.
Hand Sketching
Hand sketching is the deliberate generation of 30 to 80 small fast ugly thumbnails per concept on paper to expose structural bones before digital tools polish every lie equally.
Handoff Review Loop
Handoff Review Loop is the three-checkpoint system that kills design drift: a pre-handoff self-audit, component-first build review, and visual QA on staging that verifies the four-layer Figma file survives implementation.
Helvetica Reflex
The automatic grab for the safest geometric humanist sans of the moment, usually Inter, exactly like 1970s designers grabbed Helvetica. It is the career-safe default that feeds the sameness crisis by making every brand look interchangeable before the first pixel leaves Figma.
Heritage Brand
A brand with a long history, often spanning decades or centuries, that carries significant visual weight, tradition, and established recognition.
Hero Cell
The dominant spanning rectangle inside a bento grid that claims the most space and attention to establish clear visual hierarchy.
Hero Choreography
Hero choreography is the exact timed sequence of animations and reveals that fire in the first 800ms of a marketing site load. It turns the hero into a branded performance where every gradient bloom, type scale, and screenshot float has a deliberate cue and zero tolerance for decoration.
Hero Headline
The single opening sentence on a landing page that states exactly what the product is and who should use it.
Hero Section
The hero section is the first full-width content block on a page, built to tell a visitor where they are, what they can get, and what to do next before they decide to scroll or bail.
Hero Video
A hero video is a multi-megabyte autoplaying video that occupies the top fold of a marketing page to deliver instant motion and emotion. In 2026 it functions as the single fastest way to tank LCP, spike bounce rate, and make a premium brand read as cheap.
Hidden Product
Every great product has two products. The visible one greets new users with buttons and tours while the hidden one gives experts keystrokes palettes saved views bulk actions and power surfaces that turn the thousandth session into pure speed.
Hierarchical Grid
A hierarchical grid builds its structure from the content's reading priority instead of fixed columns or repeating modules. The most important element claims the largest space and attention while supporting pieces scale down around it.
Hierarchy Flattening
Hierarchy flattening happens when every element on a page fights for attention with the same size, contrast, color, and spacing, erasing any deliberate order so the eye has nowhere to start and users bounce in under three seconds.
Hierarchy Framework
The Hierarchy Framework is a four-step decision process that forces one primary element, ranks everything else, applies at least two levers per tier, and removes noise until the layout guides attention exactly as intended.
Hierarchy Lever
Any one of the five raw controls, size, weight, space, contrast, or motion, used to establish clear ranking between elements on a page.
Homogenization Loop
The homogenization loop is the four-node feedback cycle where designers reference the same narrow set of successful products, ship near-identical interfaces, feed them into AI training data, and receive even tighter templates that further erase variance across SaaS.
House of Brands
House of brands keeps the parent company invisible to consumers while each product runs as its own fully independent brand with separate positioning, voice, visual identity and design system.
HSL
HSL is a cylindrical color notation layered on sRGB that lets you specify colors by hue angle, saturation percent, and lightness percent. It sounded designer friendly but never delivered consistent perceived brightness across hues.
Hue Interpolation
Hue interpolation is the path colors travel between two hues during gradients and animations. OKLCH makes that path perceptually uniform so blue to yellow flows through green instead of collapsing into muddy gray.
Human Accountability (AI)
The designer's explicit ownership of every decision made in an AI-assisted creative process, ensuring a human stands behind the final output rather than deferring to the model.
Human Plane
The human plane is the visual and emotional layer of every interface that delivers flow, taste, hierarchy, and trust. In 2026 it shares every DOM with a machine plane that agents parse for selectors, names, and structure.
Humanist Sans
Humanist sans typefaces borrow proportions, open apertures, and calligraphic warmth from Roman inscriptions and oldstyle serifs instead of pure geometry or mechanical uniformity.
Hybrid Grid
A hybrid grid merges a square modular base unit with layered circular guides to control both straight elements and curves in one unified construction system. Most shipped logos live here because real marks mix hard geometry with fluid arcs that pure grids cannot handle without fighting the designer.
Hybrid Pattern
The hybrid pattern routes 90 percent of agent actions through fast tool-use APIs and falls back to computer use only for the messy long tail without clean integration points. This split delivers production economics where pure computer use bleeds cash.
I
I Like I Wish What If
I Like I Wish What If is an early stage design critique framework that collects team reactions in three buckets: what works, what needs to change, and where to explore next. It keeps feedback collaborative and prevents the ego battles that kill most reviews.
IC Role Merger
The 2026 structural collapse of separate designer and PM IC roles into one hybrid seat where the designer owns the written spec, named metric, ship date, user research, and launch write-up with Figma as a supporting appendix.
Icon Governance
Icon governance is the enforced rules, single source of truth, and contribution plus deprecation process that keeps an icon library coherent past 50 symbols instead of rotting into four competing styles by year two.
Icon Grid
The fixed pixel scaffold with live area, keylines, and optical rules that forces every icon in a library to identical size, weight, and balance no matter who draws it or when.
Icon System
An icon system is a coherent set of pictographs governed by shared rules for grid, stroke weight, optical alignment, naming, and governance. The rules let any designer or engineer extend the library without breaking the product's visual rhythm.
Identity Audit
A four-question stress test that reveals whether your SaaS has carved out a distinct visual identity or if it is simply wearing the Linear template that swallowed the industry by 2026.
Identity Image Gap
The disconnect between the brand identity you design and the brand image your audience actually perceives. It is the hidden reason many visually strong brands still struggle with conversion, recall, and trust.
In-Product Nudges
In-product nudges are contextual micro-hints anchored to the exact control a user is fighting with during live work. They drop a two-word suggestion or one-line explanation at the moment of need then vanish forever replacing the tired product tour with teaching that actually sticks.
Inbox Zero
Inbox Zero is the earned celebration that appears only after a user has cleared their queue through real effort, turning an empty state into earned validation instead of a day-one participation trophy.
Incremental Static Regeneration
Incremental Static Regeneration is Next.js's feature that rebuilds individual pages on demand or on a schedule so your marketing site stays fast and fresh without regenerating everything on every change.
Infinite Empty
The infinite empty is the single blank prompt box in AI tools that offers unlimited theoretical outputs but zero examples, templates or rails, freezing users in decision paralysis before they generate their first result.
Infinite Variant Trap
The infinite variant trap is when designers keep feeding prompts to Claude or Midjourney for one more round of options instead of picking winners and advancing through review gates. It produces 200-frame Figma graveyards, inverted five-to-one ratios, and projects that never ship.
Information Architecture
Information architecture organizes content, features, and navigation into a coherent structure so users can find what they need without confusion or rage clicks.
Inline AI
Inline AI is the pattern where AI capabilities live inside the user's existing surface as contextual actions attached directly to the object or selection they are working on.
Inline Validation
Inline validation checks each field the moment the user leaves it and surfaces specific fixes immediately instead of dumping a wall of errors on submit.
Inner Padding
Inner padding is the whitespace buffer inside each bento cell between content and borders. It targets 20 to 30 percent whitespace per cell and runs at twice the gutter width so cell sizes rank importance instead of looking like cramped or empty cards.
Input Rate
The base cost per token for an AI model to process incoming information, including system prompts, tool definitions, prior conversation history, and user input. It is the foundational price point that dictates the cost of context management in agentic workflows.
Instructional Error
Instructional error turns validation failures into clear directives that tell users exactly how to fix the problem instead of accusing them of making one.
Intent Slot
A named and typed region in a generative UI that tells the model exactly what kind of output belongs there based on user intent.
Intent-Based Naming
Naming design tokens by the specific job they perform in the interface instead of their visual appearance so both humans and machines understand the decision instead of guessing from a hex code.
Interaction States
The complete set of visual and behavioral responses for every UI component across default, hover, active, focus, disabled, loading, error, and empty conditions.
Interaction to Next Paint
Interaction to Next Paint measures the time from any user tap, click, scroll, or keypress until the browser delivers the next visual update. Target under 200ms and the interface feels alive. Cross 500ms and users register the brand as broken regardless of how clean the visuals look.
Interactive Demo
A hero pattern that embeds a functional slice of the actual product so visitors manipulate it directly and generate conviction through doing instead of reading.
Iteration Loop Speed
Iteration loop speed is the real time between typing a change request and seeing an updated live preview you can critique again in an AI app builder.
J
Journey Map
A journey map timelines every step a user takes toward a goal while documenting their thoughts emotions pain points and opportunities at each stage. It replaces assumptions with research backed insight and keeps your team focused on real user problems instead of imagined ones.
Judgment Premium
Judgment premium is the surging value of a designer's taste, strategic decisions, and ability to say no to mediocre output in a world where AI tools execute most visual and interactive work.
Just-in-Time Tutorials
Just-in-time tutorials reveal explanations for unfamiliar controls at the exact moment the user hovers, clicks or summons them instead of front-loading tours on arrival. They follow progressive disclosure so users learn inside real work exactly when curiosity peaks.
K
Kerning
Kerning is the optical adjustment of space between specific letter pairs so they sit together without gaps or collisions that the eye notices at large sizes.
Keyboard First
Keyboard first is the decision to engineer every interaction around the keyboard as primary input instead of bolting shortcuts onto a mouse-first UI. Linear built their entire product this way from day one and never softened the stance.
Keyboard Layer
The keyboard layer is the consistent, scalable system of keystrokes and letter commands that turns repeated actions into muscle memory for experts. Linear's g-then-i navigation, Figma's single-letter tools, and Superhuman's always-visible cheatsheet prove it compounds speed across the entire product.
Keyline
Keylines are the fixed geometric boundaries inside an icon grid live area that force circles squares and rectangles to sizes that deliver identical visual weight at UI scales.
Kill Criteria
Pre-written if-then rules that pause or end a design project when specific triggers are hit. They replace polite drift with mechanical decisions made while everyone is still friendly.
L
Label Design
Label design applies graphics, typography, color, and imagery to an existing container. It turns standard bottles, jars, and boxes into branded communicators without touching their physical form or materials.
Ladle
Ladle is a Vite-powered zero-config workshop that spins up component stories in under a second so design engineers can test props states and tokens in isolation without the bloat of Storybook.
Largest Contentful Paint
Largest Contentful Paint is the exact moment the biggest piece of content above the fold renders and your page finally reads as your brand. It is a Core Web Vital that designers control through every hero image, font choice, and layout decision with a 2.5 second threshold that keeps you from looking cheap.
Last Moat
The last moat is taste. When Claude Cursor v0 and Figma AI commoditize production in 2026 the only remaining edge is judgment what you reject from the model and the principles you use to defend those rejections.
Launch Write-up
The short public post-launch document written by the designer-PM covering what shipped, what moved, what did not move, and what the team learned. The artifact managers forward upward.
Layout Compression
Layout compression is the practice of folding six to eight distinct content types into one grid-based composition where cell size directly reflects real importance instead of pretending every feature matters equally.
Layout Primitive
Layout primitives are the naked divs loaded with Tailwind flex, grid, gap, padding, and alignment classes that control every bit of structure and whitespace in a React component.
Layout Scaffold
A generated starting frame that establishes section hierarchy, spacing rhythm, and content regions so designers can react instead of staring at a blank canvas.
Lazy Brutalism
Lazy brutalism is brutalist web design without the discipline. It copies the raw surfaces of 2026 sites like huge system fonts and empty space but skips the tight grids, deliberate pairings, and micro choices that make the real thing hit.
Leading
Leading is the baseline-to-baseline vertical rhythm that controls text density and directly shapes typographic hierarchy. CSS hides it behind line-height, but the values you pick decide whether headlines feel bolted down and body text feels readable or the entire system collapses into noise.
Letter Spacing
The uniform adjustment of space between all characters in a block of text. Also called tracking. Distinct from kerning, which adjusts space between specific character pairs.
Lettermark
A lettermark uses one to four characters usually the brand initials as the primary logo to compress long or awkward names into a usable visual shorthand.
LGTM Loop
The old Slack-based review habit where a designer drops a Figma link and waits for two thumbs up and a design lead glance before shipping. It worked at human speed but collapses when AI spits out eighteen thousand candidates overnight.
Lightness Drift
Lightness drift is the perceptual mismatch in sRGB and HSL palettes where colors assigned identical numeric lightness values appear at wildly different brightness levels to the human eye.
Line Height
The vertical distance between baselines of consecutive text lines. The single most impactful spacing property for readability, and the one most often set incorrectly.
Linear Aesthetic
The dark, confident visual language Linear popularized around 2022 that became the default template for ambitious SaaS marketing by 2025.
Linear Imitator
A Linear Imitator copies Linear's dark UI, command palette, minimal sidebar, gradient accents, and muted typography without copying the data model, object architecture, or constraints that made those choices work. The habit exploded in 2022 and produced years of indistinguishable SaaS products that screenshot beautifully and operate clumsily.
Liquid Glass
Liquid Glass is Apple's dynamic material system that tints itself from background content, refracts what sits behind it, and behaves like physical glass across visionOS, iOS 26 and Mac Sequoia.
Live Collaborative Demo
A hero pattern that embeds a real-time multiplayer workspace so visitors watch actual cursors, edits, and presence indicators update live on the page. The conversion trigger is the instant realization that this is not a video.
Live Composition Editor
The 2026 designer who composes interfaces by editing live code, tokens, and components directly in the running application instead of producing Figma frames. They direct AI agents like Claude Code and Cursor, ship real diffs, and own the deployed surface end to end.
Live Source of Truth
Production code and its shipped components serve as the single canonical reference for the design system. Figma becomes a rough spec surface while React primitives with Tailwind tokens and motion logic define every pixel and interaction.
LLM as Judge
The pattern of feeding AI-generated design candidates to a large language model along with a structured rubric so it returns scores, one-line reasons, and pass-fail JSON at scale.
LLM Rendered Component
A UI component selected and composed by a large language model from a fixed typed vocabulary at runtime instead of being hardcoded in a static flow.
Loading State
The UI a product displays while fetching data, running tasks, or waiting on servers. It is the most viewed screen across all user sessions yet the one most teams design last.
Local MCP Server
Local MCP Server is Figma desktop's built-in implementation of the Model Context Protocol. It runs on your machine and feeds AI coding agents the actual layer structure, auto-layout values, component instances, and design tokens from your files instead of forcing them to guess from screenshots.
Lockup Grammar
Lockup grammar is the codified rule set that dictates exactly how a parent brand mark relates to sub-brand marks in scale, spacing, alignment, color, and hierarchy across every context. It turns endorsed brand architecture into a repeatable system instead of a recurring argument.
Logo Audit
A logo audit is the exhaustive documentation of every current logo variant, its real deployments, technical performance, customer perception, and competitive context before any redesign work starts.
Logo Brief
The single-page contract that forces the client to answer five specific questions on business, audience, competitors to avoid, primary surfaces, and brand adjectives with anti-adjectives before any sketching starts.
Logo File Inventory
The complete set of 15-20 precisely named files that turn a responsive logo system into something developers, printers, and marketers can actually use without guessing or breaking the design.
Logo Grid
A logo grid is a system of geometric guides using squares, circles, ratios, and angles that locks proportions, curves, and spacing so the mark stays consistent across sizes and handoffs.
Logo Kit
A logo kit is the structured folder containing every file format, variant, color space, and a README so developers, printers, and marketers grab the exact asset they need without guesswork or amateur mistakes.
Logo Ladder
The Logo Ladder is the four-tier system of full lockup, primary mark, simplified mark, and monogram that keeps a brand legible and authoritative from billboards to 16-pixel favicons.
Logo Loader
A logo loader is a short seamless loop that brands app loading states, data fetches, and initial renders. It runs 1-2 seconds max, ships as lightweight Lottie under 50KB, reads at 24px, and collapses to a rock-solid still frame the instant motion ends.
Logo Lockup
A logo lockup is the fixed spatial relationship and exact measurements between symbol, wordmark, and clear space that turns a loose idea into a repeatable unit.
Logo Refinement
Logo refinement is the post-exploration phase where you lock one direction and surgically adjust every curve, optical balance, spatial gap and grid alignment so the mark performs at every size and context without losing the strategy it was built to serve.
Logo Simplification Rules
Four rules for reducing logos at small scales: drop hairlines below 1px, open counters to prevent filling, simplify bezier curves, and convert outlines to solid mass for better rendering.
Logo Sting
A logo sting is the 1-3 second audio mark that identifies your brand instantly with the screen off.
Logo System
A logo system pairs a primary wordmark with a lettermark or monogram plus the strict rules that dictate when each version ships so the brand stays sharp from 16 pixel favicons to highway billboards.
Logo Tier
A logo tier is one of four distinct purpose-built variants in a responsive logo system, each engineered for a specific size and context ranging from full wordmarks at 600px to stripped icons that survive at 16px.
Logo Wall
A logo wall is a curated grid or strip of recognizable company logos placed on a landing page to deliver instant structural social proof and transfer credibility before features or testimonials appear.
Logomark
A symbol or icon that represents a brand without any text. The Apple apple, the Nike swoosh, the Airbnb Belo.
Logotype
A logo made entirely of styled text, where the brand name IS the logo. Google, Coca-Cola, and FedEx are logotypes.
Long Horizon Capability
The ability of an AI model to sustain coherent reasoning, self-correction, and output quality across multi-hour, multi-thousand-decision tasks that cause earlier models to drift, contradict themselves, or collapse.
Long-Horizon Agent
A long-horizon agent grinds through complex multi-step tasks for hours while holding its goals, decisions, and context without drift or repetition. Claude 4.7 stabilized these agents for production by flattening the failure curve that wrecked 4.6 models past the ninety-minute mark.
Lossy compaction
Lossy compaction is the process of reducing an LLM's context window by summarizing its contents, intentionally discarding granular detail to free up token space and maintain conversational flow.
Lottie
Lottie is a JSON-based animation format that renders After Effects vector animations natively across web, iOS, and Android. Airbnb open-sourced it in 2017 to end the tradeoff between rich motion and small file size, letting a motion designer's timeline ship as lightweight data instead of a heavy video or a bloated GIF.
M
Machine Readable Structure
Machine readable structure is the collection of semantic HTML, JSON-LD, AGENTS.md, llms.txt, and MCP tool definitions that agents parse to build an accurate model of your product before they act.
Machine-Readable Design System
A design system structured with named tokens, single-source components with explicit variants, and zero raw values so AI tools can read and compile from it directly instead of guessing at pixels.
Macro Whitespace
Macro whitespace is the deliberate large-scale spacing between major page sections, wide margins, and air around heroes that lets the eye chunk content into clear regions before reading a word.
Magnetic Cursor
A magnetic cursor pulls the pointer toward a primary CTA when it enters an 80 to 120 pixel radius, using real-time mouse tracking and smooth lerping to create a gravity effect that signals importance and previews the click.
Maintenance Bill
The compounding hourly tax of keeping design tokens, components, docs, and live products in sync after the initial build phase.
Major Third
Major third is a 1.25 modular scale ratio that multiplies a 16px base to generate every size in a tight typographic ladder ideal for dense interfaces.
Malleable Software
Software designed so users can reshape it to fit their exact needs. The people using the tool become its editors instead of passive consumers locked into rigid templates.
Manual Bridging
Manual bridging is the repetitive tax designers pay copying screenshots, pasting context from Figma Notion and GitHub, then describing their work in text because AI cannot see the actual files.
Manuscript Grid
A manuscript grid is a single column of text optimized for long form reading with a deliberate 65 to 75 character measure and generous whitespace margins that remove every competing element.
Mark Only
The standalone symbol from a logo system with zero text attached. It proves the mark can carry the full brand weight on product UI, merch, favicons, and motion stings without constant redesigns.
Mascot
A mascot is a named character that serves as the public face and personality of a brand, delivering warmth and storytelling no static mark can touch.
Mask Animation
Mask animation reveals a logo by sliding or expanding a shape that uncovers the mark like an aperture in motion. All the complexity lives in the mask path leaving the logo vectors untouched which makes the pattern portable across Lottie WebM MP4 and every size from 24px loaders to 4k video.
Masonry Layout
Masonry layout is an algorithmic grid that packs content blocks of uneven heights into columns by dropping each new item into the shortest column available creating a staggered brick wall effect with zero wasted space.
Material Palette
The selection of physical materials (paper, board, glass, aluminum, film) used in packaging design. Material choice communicates brand positioning before any graphic element.
Material Parallax
Material parallax moves UI layers at different speeds according to their z-depth so the foreground races ahead while background elements lag creating instant physical depth on flat screens.
Material Signals
Material signals are the instant expectations customers form from a package's physical substance before they process any text or graphics. Kraft says artisan and honest. Thick glass says ritual and premium. These choices shape perception faster than any logo or color block.
MCP Config
The JSON file that tells Claude Desktop and Cursor exactly which MCP servers to launch, what arguments to pass, and which tokens to use so your AI can read live Figma files, audit GitHub repos, and update Linear tickets without you copying screenshots or re-briefing context every session.
MCP Host
MCP Host is the client application like Claude Desktop or Cursor that implements the Model Context Protocol to discover servers, dispatch tool calls, and feed real data back to the AI model so it stops guessing at your designs.
MCP Prompt
MCP Prompt is a reusable template hosted on an MCP server that pulls live resources from tools like Figma Linear and Notion then runs a structured workflow with strict output rules so teams stop rewriting the same instructions every sprint.
MCP Resource
Read-only data an MCP server exposes so AI models can access real structured context from Figma files, Notion databases, or GitHub repos instead of relying on screenshots or copy-pasted specs.
MCP Server
MCP Server is a local process that implements the Model Context Protocol so AI editors like Cursor can read live data from Figma including exact tokens component structures and layout values instead of guessing from screenshots.
MCP Tool
An MCP Tool is an action an AI model invokes through a standardized MCP server. It lets Claude or Cursor call real operations like reading exact Figma frame coordinates, creating a Linear issue with linked PRs, querying scoped Postgres tables, or updating a Stripe subscription with live state management and scoped auth.
MCP Transport
MCP Transport is the three connection methods (stdio, HTTP+SSE, and WebSockets) that carry JSON-RPC messages between AI clients and MCP servers. Same server code runs everywhere. The choice determines whether your agent stays a laptop toy or actually ships to a team.
MCP Workspace
A live collection of MCP servers that turns Figma, Notion, design tokens, Linear, and data sources into one continuous context for AI agents like Claude Code. The new design surface that replaced screenshots and copy-paste.
Meaning Gap
A meaning gap is a polished logo that communicates nothing specific about the brand. It passes form and scale tests yet a stranger cannot describe the company in one clear sentence after seeing it.
Mega Menu
A wide multi-column overlay triggered from a top navigation link that surfaces deep hierarchies, grouped categories, sublinks, and visual tiles in one scannable panel.
Memory Card
A memory card is the atomic unit of stored memory in an agent system. It holds one discrete fact, preference, or behavior signal plus metadata for timestamp, scope, source, and expiration so the agent can surface, edit, and expire it without turning into a creepy black box.
Memory Chip
A visible, editable, and removable UI element on the prompt surface that shows exactly what context the model retains for the current session.
Memory Files
Memory files are persistent storage within Claude Code, like CLAUDE.md, designed to hold invariant rules, project-specific facts, and core instructions that survive conversation resets and context compaction. They ensure critical information about your project or design practice is always available to the AI, reducing repetitive prompting.
Memory Inspector
The memory inspector is the full-screen control center that makes every stored preference, fact, and behavior signal visible, editable, and exportable.
Mental Model
The user's working theory of how an AI system behaves, what it can and cannot do, how to speak to it, and what success looks like. It must be deliberately installed in the first sixty seconds because the single prompt bar hides every clue.
Mess Multiplier
A mess multiplier is a sloppy design system that AI tools treat as truth, turning every conflicting button variant, unnamed hex, and missing state into consistent, high-speed garbage across every generated screen.
Metric Carousel
A metric carousel is a horizontal strip of KPI cards each showing one big number, one colored percentage change, and one microscopic sparkline. It creates the illusion of comprehensive oversight while delivering zero priority, zero context, and zero tied actions.
Metric Matched Fallback
A metric matched fallback is a system font declared in @font-face with ascent-override, descent-override, line-gap-override, and size-adjust values tuned to occupy the exact same space as your variable font. Pair it with font-display: swap and your text renders instantly with zero layout shift when the real font arrives.
Micro Whitespace
Micro whitespace is the precise spacing between letters, lines, labels, inputs, icons, and UI controls that determines whether dense interfaces feel calm and premium or cramped and sloppy.
Micro-Interaction
Micro-interactions are precise engineered moments of motion that direct user attention, signal interactivity, and reduce decision friction on 2026 interfaces.
Migration Wall
The migration wall is the hard limit where your stack's assumptions collapse under real business growth, forcing a costly rebuild of layouts, content models, and team workflows. Framer walls at CMS complexity around 50 pages, Webflow at pricing cliffs and custom logic needs, Next.js almost never if you keep engineers on payroll.
Mini-App Shell
A thin launcher frame that summons self-contained mini-applications for specific tasks instead of navigating through a single monolithic interface with persistent navigation. Each surface owns its own UI completely while sharing only the account context and data layer.
Minimalist Logo Design
The practice of reducing a logo mark to its most essential visual elements to maximize recognition and meaning.
Minimum Viable Visual System
The three layer stack of design tokens, UI primitives, and taste that delivers consistent interfaces for the 95 percent of teams who should never build a full design system.
Missing Tokens
Missing tokens refers to the failure to translate design system specifications, particularly for typography, into code-ready design tokens or CSS custom properties. This disconnect forces developers to guess values, leading to inconsistencies and design drift between design files and the live product.
Mobile Desktop Parity
Mobile desktop parity is the principle that mobile and desktop versions of a site must deliver the same information, offers, proof, and speed. Layouts adapt but nothing important gets hidden or slowed down.
Mobile Input Modes
Mobile Input Modes are the HTML type and inputmode attributes that tell the browser exactly which virtual keyboard to show users on phones and tablets.
Mock Polishing
Mock polishing is the senior designer practice of taking junior roughs and refining them into pixel-perfect design-system-compliant mocks with perfect hierarchy, tokens, states, and brand fidelity. AI tools like Figma AI, v0, Galileo, Magician, and Subframe ate ninety percent of this work in 2025 shifting the bottleneck from creation to selection.
Mockup Deliverable
The polished Figma frame that designers shipped as the final approved artifact for fifteen years before running code took its place in 2026.
Model Context Protocol
An open standard introduced by Anthropic that lets AI agents read and interact with external tools, data sources, and services through a shared interface.
Model Lock-in
Model lock-in traps teams when an AI code editor ties them to one model family with no BYOK or clean switching. Six months later the model drifts on their codebase and velocity collapses with no escape hatch.
Modular Grid
A modular grid carves the canvas into uniform cells on both axes so every headline, card, image, and metric snaps to the same invisible graph paper. Density becomes a lever instead of an accident.
Modular Scale
A modular scale generates harmonious type sizes by multiplying a base value by a consistent ratio like 1.25 or 1.618, replacing guesswork with predictable hierarchy that adapts across devices.
Modular Type Scale
A modular type scale applies one ratio to one base size to generate every font size in a product so every heading body and caption shares the same mathematical DNA instead of random pixel values.
Mono-Material Construction
Mono-material construction uses one material family for every element of a package so the whole thing recycles in a single stream without disassembly or contamination.
Monochromatic
A color scheme built from a single base hue, achieving depth and variation through its tints and shades.
Monogram
A monogram interlocks brand initials into one dense geometric mark by forcing their strokes to cross and share edges. It serves as the compact status tool in a logo system after the wordmark and lettermark have been decided.
Mood Board
A curated visual collection of images, textures, colors, typography, and references used to define a brand's aesthetic direction before detailed design begins.
Most Popular Badge
The Most Popular badge is a small label that marks the pricing tier chosen by the majority of actual paying users. It speeds up decisions when tied to cohort data and becomes obvious theater when painted on the tier with the best margin.
Mother Color
A technique where a small amount of one base color is mixed into every other color in a palette, creating a subtle shared undertone that unifies the entire scheme.
Motion as Information
Motion as information means every animation on your interface must communicate a specific state change, user action, or content update or it gets cut. It turns movement from decoration into a functional language that reduces errors and builds trust.
Motion Brief
A motion brief is the one page document that defines every constraint for an animated logo project before a single frame is created. It replaces vague direction with specific modes, lengths, references, and success criteria.
Motion Budget
A motion budget is the hard cap on how many elements can move, for how long, and at what performance cost so your interface stays fast, calm, and focused. It limits every animation to GPU-friendly transform and opacity properties, caps concurrent movement at one primary action, and demands every transition earn its place against CPU, battery, and user attention.
Motion Control Prompt
A motion control prompt forces AI video models to execute exact single-axis camera orbits with zero improvisation so you can extract clean parallax frames for interactive Figma prototypes.
Motion Design
Motion design is the behavior layer of a product that controls how elements enter, leave, respond, and relate over time so users instantly understand state changes without extra copy.
Motion Easing
Motion easing is the acceleration and deceleration curve that dictates how interface elements move over time. It turns mechanical linear tweens into motion that feels human, using three core curves to communicate state changes without stealing focus.
Motion Lever
Motion lever uses targeted animation and interaction to seize attention and finish the read sequence after size, weight, space, and contrast have set the stage.
Motion Modes
Motion modes are the four standardized contexts every animated logo must serve: loader, intro, loop, and transition. They force teams to build a flexible system with shared rules instead of one hero clip that only works on a showreel.
Motion Sting
A 2-to-4 second animated logo sequence built as the video-native variant in a complete brand mark family. It expresses the brand's personality through deliberate timing, easing, and sequencing instead of generic spins or templates.
Motion System
A motion system is a set of three to five named duration and easing tokens that enforce consistent animation behavior across every screen so the product never feels like it was designed by committee.
Motion Taste
The ability to specify, implement and defend micro-interactions, easing curves, durations and animation purpose so a product feels built instead of assembled.
Motivational Family
One of four groups that cluster the 12 brand archetypes by the core human motivation they address: stability, belonging, mastery, and freedom.
Multi-Agent Workflow
An AI setup where multiple specialized agents collaborate on a single task, each taking responsibility for planning, execution, critique, or revision instead of relying on one model in one pass.
Multi-File Refactor
The capability of an AI code editor to identify, update, and verify changes across many files at once while preserving system consistency.
Multi-Server Orchestration
The client coordination layer that spins up multiple MCP servers inside one AI session, discovers their resources and tools through JSON-RPC, routes calls, merges context from Figma Linear Postgres and GitHub, and keeps the full workspace coherent across transports.
Mythos Class
Mythos Class is Anthropic's top capability tier above Opus, powering Fable 5 for general users behind heavy classifiers and Mythos 5 for vetted researchers with several guardrails removed.
N
Named Metric
A named metric is the single specific number a designer-PM owns before they open Figma. It turns every ship into an accountable bet on activation rate, conversion lift, retention curve, latency, or NPS instead of subjective praise at design review.
Named Variant
A single master component with explicit properties like variant=primary and state=hover that a machine can read and resolve without guessing. It replaces scattered duplicates with descriptive names that force AI tools to play detective with your intent.
Naming Gauntlet
The naming gauntlet is the eight-filter pipeline that stress-tests every brand name candidate for trademark strength, practical viability, cultural safety, and twenty-year durability before emotional attachment forms.
Narrow Agent
A narrow agent is an AI system built for one specific workflow with fixed steps, clear success criteria, supervised execution, and instant handoff to a human at the first sign of trouble.
Nearest Frame Selection
Nearest frame selection calculates cursor position on mousemove, determines the dominant axis, then displays the single pre-rendered frame that matches closest. It creates convincing parallax using discrete stills instead of blending or video.
Negative Prompt
A negative prompt is the explicit list of visual cliches, technical errors, and aesthetic mistakes you ban from every AI-generated image to stop your brand from dissolving into generic slop.
Negative Space
The empty area around, between, and within design elements. In logo design, negative space is an active compositional tool, not leftover blank area.
Neo-Grotesk
Neo-Grotesk is a sans-serif category defined by even stroke weight, closed apertures, tall x-heights, and mechanical neutrality that projects precision and control without any decorative noise.
Neumorphism
Neumorphism is the short lived 2020 trend that extruded soft plastic like surfaces using inset and outset shadows but delivered zero semantic meaning or accessibility.
Neutral Scale
The 9-to-12-step ladder of tuned grays that handles surfaces, text, borders, backgrounds, and contrast in any real interface. Built before the anchor is finalized and adjusted to its temperature, it turns a moodboard into infrastructure.
Never Say List
The never say list is a hard coded roster of banned words phrases and constructions paired with their exact on brand replacements. It stops corporate sludge from diluting your voice and gives every writer and AI generator unambiguous rules instead of vague warnings.
Ninety Day Rule
The Ninety Day Rule audits your Figma plugins by forcing three questions: have you opened it in the last 90 days, does it solve a real problem more than once a month, and does it do work Figma cannot handle natively in 2026. Fail any question and delete it immediately.
Ninety-Second Skim
The ninety-second skim is the 90-second filter every hiring manager runs on your portfolio link in 2026. They scan for one shipped product with real users, one decision log, and one piece of public writing. All three fire and you advance. Anything missing and the tab closes before your hero image loads.
No Onboarding
No onboarding is the deliberate elimination of welcome modals, product tours, and gating checklists so users reach real output in under thirty seconds while the product itself does the teaching.
Noisy Gradient
A noisy gradient layers subtle film grain or procedural static over a smooth OKLCH color transition to kill banding, add tactile depth, and create premium texture that ships in under 4kb and paints instantly.
Non-Text Contrast
Non-text contrast is the WCAG 2.2 rule that requires buttons, icons, form borders, focus rings and any meaningful graphic to hit at least 3:1 against their background.
O
Object First Naming
Object first naming labels icons by the literal object they depict instead of the action or metaphor they represent. A magnifying glass is search or magnifier never find or zoom because objects stay fixed while contexts shift.
OKLCH
OKLCH is a CSS color space built on OKLab that delivers perceptually uniform lightness, predictable hue shifts, and native wide-gamut support for P3 and Rec2020.
On-Color
The Material 3 pattern that pairs every base color with a guaranteed accessible counterpart so designers pick roles instead of fighting ratios.
One Column Form
The rule that every form must stack all fields in a single vertical column to match natural reading flow and eliminate the micro decisions that kill completion rates.
One Idea Per Mark
A principle of minimalist logo design where the mark communicates a single, clear concept without trying to convey multiple messages simultaneously.
One Line Positioning
The single sentence on your portfolio homepage that tells hiring managers exactly what kind of work you do and who you do it for so the right people stop and the wrong ones move on.
One Metric That Matters
One Metric That Matters is the single measurable outcome written into a design brief that decides whether the project succeeded. It forces every decision, deliverable, and deadline to serve one observable number instead of a pile of feelings or stakeholder opinions.
One Recommended Direction
One recommended direction is the practice of presenting a single fully resolved brand identity system instead of multiple concepts so clients confirm strategy they already signed off on rather than picking favorites based on taste.
One Second Hero
The one second hero is the opening section of a SaaS landing page engineered to stop bounces instantly with a benefit-first headline, a precise subheadline naming the audience and outcome, and one unmistakable CTA.
One Voice Per Role
The rule that every typeface owns exactly one job, heading or body or UI, with zero overlap so your system stays sharp instead of muddy.
One-Color Test
A crucial design test where a logo is evaluated for its strength and recognition when rendered in a single color, typically black on white.
One-Line Outcome
The opening sentence of a case study that states exactly what shipped and what business or user number it moved, written like a headline instead of creative writing.
One-Page Constraints Sheet
The single signed page that replaces 30-slide decks by locking in six decisions: one named decider, a screenshotable done state, an explicit out-of-scope list, the exact budget kill number, the calendar kill date, and the one metric that matters.
One-Shot App
A complete, functional software product, often including user authentication, payment processing, and real data, built and shipped within a single extended work session using AI-augmented tools.
Opinionated Default
An opinionated default is the product team's explicit bet about what the user should see and do first. It replaces neutral widget grids with a single prioritized question, a product-sorted list, and clear actions, forcing focus where it matters most.
Optical Correction
Adjusting perfect geometry so it looks correct to human eyes. Overshoot rounds by 1 to 2 percent, thicken verticals, shift intersections. The invisible craft that separates amateur marks from professional ones.
Optical Size
Optical size is a variable font axis that redraws letterforms for specific rendering sizes, delivering sturdy readable shapes at 11px and refined high-contrast forms at 72px from the same file.
Optical Size Axis
The variable font axis that modifies contrast, x-height, and detail based on rendered size so one file can serve both body text and display headlines correctly.
Optimistic UI
Interface behavior that applies user actions instantly as if they succeeded, then reconciles with the server in the background and rolls back gracefully on failure.
Orphan Weight
An orphan weight is a one-off use of an extreme font weight like Thin 100 or Black 900 that appears exactly once in your entire system with no matching usage anywhere else.
Output Ownership
Output ownership measures whether an AI app builder hands you a portable repo you control forever or locks you into its hosted runtime with painful migration costs.
Output Spec
Output Spec is the final section of a structured prompt that dictates the exact technical requirements for the deliverable including dimensions file formats code standards component variants and naming conventions.
Output Transformation
The mandatory process of rebuilding, refining, and customizing AI-generated output until it is genuinely unique, aligns with brand systems, and cannot be reproduced by a generic prompt.
Owned Color
Owned color is a specific hue a brand selects then deploys with fanatical consistency until customers recognize the company by that shade before the logo appears.
P
Package Pricing
Package pricing presents clients with three fixed-fee options, each tied to a clearly differentiated scope, so they select a tier instead of negotiating your single number downward.
Packaging Identity
The complete system of structural, material, and visual decisions that make a product recognizable across shelf, e-commerce, unboxing, and social media contexts.
Packaging Typography
Packaging typography sets strict hierarchies, loose tracking, and font choices that perform from three feet away on shelf before revealing detail up close. It treats type as a physical object that must fight for attention against color, shape, and competitors in three seconds or less.
Palette Layers
Palette layers split a brand palette into five functional systems: anchor, neutrals, semantic, state, and dark. Each layer owns its job so the colors survive real product work instead of collapsing at the first disabled button or dark mode request.
Parallax Hero
A cursor-reactive cinematic header built in Figma with AI-generated orbital frames that shift perspective based on mouse position to create convincing depth without 3D models, shaders, or developers.
Parallel System
A parallel system is an independent dark mode palette designed from scratch instead of inverting or tweaking your light mode colors. It reuses the same semantic token names but maps them to primitives tuned for how eyes actually behave against near-black surfaces.
Partial Prerendering
Partial prerendering is a Next.js rendering model that ships a static HTML shell from the edge instantly while streaming in only the dynamic pieces at request time.
Passive Negative Space
Passive negative space is the intentional margins, padding, gutters and leading that keep interface elements from suffocating each other. It creates breathing room so users can parse complex screens without cognitive overload.
Path-Scoped Rules
Path-scoped rules are localized instructions defined in subdirectory CLAUDE.md files, activating only when Claude's context window includes files from that specific path. They allow Claude to dynamically adapt its behavior and coding standards based on the exact part of the codebase being edited.
Pattern Recombination
Pattern Recombination extracts specific design patterns from three to five product screenshots then blends them through targeted AI prompts to create original interfaces that feel both fresh and familiar.
PDF embeds vector logo data that opens on any device without special software. It serves as the universal handoff format for print production and vendor deliverables.
Per Job Cost
Per job cost measures the true expense of finishing one complete AI task. It multiplies token price by actual usage then adds retries, quality drift, human rework time, and latency tax instead of trusting the sticker price per million tokens.
Perceived Craft
Perceived craft is the immediate impression that a logo was engineered with rigorous thought and precision. It transforms negative space from empty area into strategic communication that makes the entire brand feel smarter.
Perceived Performance
Perceived performance is the gap between actual latency and how fast the product feels to the user. It lives in every loading state, skeleton, optimistic update, and branded transition that either builds trust or quietly erodes it.
Perceived Speed
Perceived speed is how fast an AI product feels to the user. It depends on feedback density and useful signals during the wait rather than raw milliseconds or total response time.
Perceptual Uniformity
Perceptual uniformity means color scales where each numerical step creates equal visual difference to the human eye instead of equal math steps in broken models like HSL. Systems built on OKLCH or LCH keep contrast stable when you swap hues or spin up new themes.
Perfection Paralysis
Perfection paralysis is when design system teams obsess over solving every edge case and hypothetical scenario before launch causing the entire project to stall indefinitely while the product ships inconsistent interfaces every single day.
Performance as Aesthetic
Performance as aesthetic is the 2026 design principle where load speed, technical restraint, and lean code become the primary signals of premium quality. The fastest sites now feel the most trustworthy because every millisecond reads as confidence.
Performance Budget
The one-page contract signed by design, brand, and marketing that locks LCP under 2.5s, JS under 200KB, fonts to two weights max, and every other speed target before the first wireframe leaves Figma.
Persistent Sidebar
Persistent sidebars lock navigation into a fixed left column that remains visible across all pages and scroll positions. The spatial permanence builds user confidence in deep products where constant reorientation would otherwise create friction.
Personal Software
Software built by one person for themselves plus a handful of specific people they know by name. It delivers uncomfortable precision for a tiny audience instead of generic features for strangers.
Pictorial Logo
A pictorial logo is a symbolic image or illustration that functions as the primary brand mark with zero reliance on letters or words. Nike's swoosh, Apple's bitten apple, Mercedes' three-pointed star, and the WWF panda all bet the entire identity on one visual metaphor that triggers instant recognition.
Pictorial Mark
A pictorial mark is a simple drawing of a real-world object that instantly communicates the brand without needing the name attached.
Pixel Level Decisions
Pixel level decisions are the 1px to 4px spatial and optical adjustments that separate mechanical correctness from felt quality in interfaces.
Plan Surface
The structured editable list that shows every step an agent plans to take before it executes a single action. It turns invisible model reasoning into a reviewable contract the user can modify inline or reject outright.
PM With Taste
The senior IC designer in 2026 who owns the full product cycle like a PM while applying sharp taste to every decision. They write the spec, name the metric, lock the ship date, run the research, and publish the writeup instead of waiting for someone else to do it.
PNG
PNG is a lossless raster format that stores logos as fixed pixel grids with transparency support making it the universal digital fallback yet it pixelates on scale and destroys print jobs.
Port Stack
The four artifacts (shipped product, components, decision logs, motion demo) that get you hired in 2026. Proof over polish.
Portfolio Bloat
Portfolio bloat is every irrelevant artifact and self indulgent paragraph that pads case studies with client histories, mood boards, ideation grids, and personal journeys while burying the outcomes, constraints, decisions, and numbers that actually get you hired.
Portfolio Worship
Portfolio worship is the belief that a beautifully presented case study collection equals actual design skill. It turns the artifact into the work and leaves designers unequipped for real product decisions.
Portmanteau Name
A portmanteau name fuses fragments of two words into one new term that hints at both roots while standing as a single ownable identifier.
Positioning Statement
The one sentence that defines exactly who your brand serves, what category it owns, how it differs, and why anyone should believe it.
Post-Clear Empty
The empty state that appears after a user intentionally empties a list, completes tasks, or archives content. It represents accomplishment rather than absence and should celebrate without overdoing it.
Post-Dashboard
The family of focused subtractive patterns that replace the 2015 chart-grid dashboard with surfaces built for action instead of renewal demos.
Post-Hoc Rationalization
Post-hoc rationalization is writing decision logs after launch metrics land so every choice looks inevitable and every outcome looks like a win. It replaces real alternatives and live constraints with theater that senior reviewers spot in the first paragraph.
Poster Pairing
Poster pairing is two fonts that look perfect on a specimen or moodboard but collapse the moment you drop them into real landing pages, longform articles, or product UIs.
Power Surfaces
Power surfaces are expert-only interfaces that live one keystroke away and ignore new users completely. They deliver raw speed and density for the cohort that opens your product twenty times a day.
Power User UX
Power user UX is the hidden layer of interfaces built for experts who open your product twenty times a day and expect every action to bend to their speed instead of the other way around.
Prefers Reduced Motion
Prefers reduced motion is an operating system setting that signals users want less animation usually for vestibular migraine or attention reasons. Respecting it is now table stakes for accessible products.
Preview Deployment
A temporary live URL that spins up for every pull request or branch so designers can click through real code with test data minutes after the engineer pushes.
Price Honesty
Price honesty displays your exact self-serve pricing tiers with real numbers, inclusions, and upgrade buttons directly on the landing page so buyers can evaluate cost without forms, calls, or gates.
Pricing Anchor
A pricing anchor is the first high number you drop in a sales conversation to reset client expectations upward. Lead with your top package or rate band so your target middle price feels like the reasonable choice instead of a stretch.
Pricing Calculator
A pricing calculator is an interactive tool on a pricing page that replaces static tiers with sliders inputs and toggles so visitors can model their own usage and see a live total. It only works when the product's core unit is inherently variable computation like transactions or API calls.
Pricing Page
The page that forces a product team to commit to specific prices, target users, and value metrics instead of hiding behind vibes and feature lists.
Pricing Worksheet
The one-page internal tool that calculates your quote floor from hours and rate, pulls your target from the 2026 rate bands, sets a walk-away number, and builds three differentiated packages before any proposal is written.
Primary Lockup
The primary lockup is the default pairing of symbol and wordmark that owns the hero surfaces: website headers, pitch deck covers, and business cards.
Primary Mark
The primary mark pairs the brand symbol with the wordmark sans tagline and owns the 80 to 300 pixel range where most people encounter the brand. It is the system's foundation. Every spatial decision made here, from mark-to-text ratio to optical kerning, gets inherited by the simplified mark, monogram, and full lockup.
Primitive Layer
The primitive layer holds raw unchanging values like hex colors, pixel spacing, and radii that form the foundation of a token system. Every semantic and component token references these values but never applies them directly.
Primitive Token
Primitive tokens are the raw unchanging values at the base of any serious design system. They store exact hex codes, pixel measurements, millisecond timings and RGB values that semantic and component tokens reference but never override.
Print Canon
The accumulated principles, techniques, and landmark projects from a century of print design that digital teams keep reinventing from scratch.
Pro canvas
A professional-grade design environment built for precision, collaboration, and the full lifecycle of product development from detailed refinement to production handoff. It is where designers transform initial concepts into shippable, accessible, and scalable user interfaces.
Problem Framing
Problem Framing is the critical upfront work of defining and structuring a design challenge, ensuring you are solving the right problem for the right people, before any solution is attempted. It is the strategic act of identifying the core issue, its boundaries, and its impact, setting the stage for effective design.
Problem Reframe
The section after the proof strip that names the visitor's exact daily friction in their own words so they stop scrolling and think this page gets me. It creates instant recognition that qualifies the right buyers and quietly bounces everyone else.
Process Theater
Process theater is the performative display of design activities that offer little genuine contribution to the final product, often used to create the illusion of a rigorous process. It is a smokescreen, obscuring a lack of genuine insight or shipping ability, and is now a liability in modern portfolios.
Product Demo Block
The product demo block shows your SaaS tool performing its core job in real conditions so visitors watch it work instead of imagining whether it might.
Product Designer
A product designer owns both UX decision architecture and UI visual execution end to end for a specific product area or feature.
Product Hardening
The deliberate audit and fix process that closes security, brand, UX, and structural gaps left by AI tools so a fast demo becomes a real business that survives scrutiny and scale.
Product Line Extension
A product line extension adds variants or categories to a master brand using plain descriptors that never escape the branded house. Descriptors like MacBook Pro, Oatly Barista Edition, or Air Max 270 keep one design system, one voice, and one equity pool intact instead of creating sub-brands.
Product RFC
A Product RFC is the actual decision document a team writes to resolve hard tradeoffs before any code ships, then publishes openly as a first-class product artifact instead of hiding it in private wikis.
Product Shot Centered
A hero pattern where a beautifully photographed physical product dominates the viewport at massive scale, supported only by a short declarative headline and one CTA. It bets the entire conversion on whether the object looks desirable enough to stop the scroll by itself.
Product-adjacent Marketing
Product-adjacent marketing is when your marketing site shares the exact same codebase, component library, design tokens, and data layer as the product it sells. In 2026 this usually means a Next.js monorepo where marketing pages import real production components instead of faking them in a visual builder.
Production Assistant
The new bottom rung on the post-2025 design ladder. Production assistants run Figma AI, v0 and Claude to generate screens and flows then perform quality control and ship output to exact spec.
Production Environment
The live environment your customers actually use, running on real data, real money, and real consequences where mistakes are immediately public.
Production Layer
The bundle of tasks involving actual making of interfaces, components, visuals, motion, and variants that AI collapsed in cost and speed by 2026.
Production-ready design
Design output that meets all technical, accessibility, and quality standards for immediate implementation by engineers and shipping to users. It is the comprehensive, error-free blueprint for a live product.
Productized Design Service
A fixed monthly subscription where clients pay one price for a tightly defined scope of design work delivered through a request queue on a predictable turnaround.
Productized Service
A productized service is a fixed-scope fixed-price design deliverable with a locked turnaround that pairs senior taste with AI production. In 2026 solo designers sell brand systems with prompt packs at twenty-two thousand dollars landing page sprints at twelve to fifteen thousand and design system kits at twenty-five to thirty thousand each shipping in two to four weeks at seventy to eighty-five percent margin.
Progress As Content
Progress as content turns the loading state into the main event. Users read live logs, reasoning traces, build steps, and diagnostic output as the actual product instead of staring at a spinner while the final result cooks.
Progress Stream
The progress stream is the real-time structured log of every tool call, file edit, command output, and state change an agent makes so the user can supervise instead of guess.
Progressive Disclosure
An interface pattern that shows the minimum information needed for the current decision, then reveals additional detail only when the user signals they want more.
Project Pricing
Project pricing locks a fixed fee to a scoped deliverable so AI efficiency improves your margin instead of cutting your invoice. In 2026 it became the default model for brand systems, token libraries, and MCP integrations that ship prompt packs clients can actually use.
Prompt
The input text, question, or instruction given to an AI model to generate a response. The quality of the prompt directly shapes the quality of the output.
Prompt Anatomy
Prompt anatomy is the fixed five-part structure every production prompt must follow: system, scope, examples, constraints, and output format. It turns disposable strings into versioned, reusable components that survive model updates and team scaling.
Prompt as Component
Prompt as component turns system prompts into versioned reusable primitives with strict contracts for inputs outputs and tool use so they can be composed into reliable agent systems like code modules.
Prompt As Input
Prompt as input replaces rigid forms, dropdowns, and wizards with natural language so the model interprets intent and handles structuring. Cursor Cmd-K, Linear command bar, and Perplexity search all run on this pattern making the AI the primary surface instead of an accessory.
Prompt Caching
Prompt caching freezes repeated context like system prompts, codebases, and rubrics on Anthropic servers so you pay full input price once then hit a thirty percent cheaper read rate for up to one hour on every follow up call.
Prompt Component
A prompt component is a reusable scoped instruction unit for AI models built with the same discipline as a UI component including anatomy variants versioning evals and distribution.
Prompt Constraints
Prompt constraints are the explicit rules and exclusions you embed in an AI prompt to block generic patterns and enforce your exact standards. They stop the model from defaulting to stock slop and force output that matches the quality bar you set for human collaborators.
Prompt Context
Prompt context tells the model exactly what the output is for and who will see it. It keeps the AI from solving the wrong problem for the wrong audience.
Prompt Drift
Prompt drift is the slow degradation that happens when designers rewrite the same prompt from memory or stale notes week after week. Each version loses precision until outputs that once delivered senior-level rigor read like vague intern drafts.
Prompt Engineering
The practice of writing instructions that produce consistent, usable output from a language model. Functionally identical to writing a good creative brief.
Prompt Librarian
The prompt librarian owns the design team's prompt library the same way a design systems engineer owns components. They curate every prompt with five-part anatomy, enforce variants and semver, run evals on every change, maintain rubrics, and connect real conversion data back into the system so prompts sharpen over time instead of rotting.
Prompt Library
A prompt library is a git-backed, versioned collection of prompts structured like a design system. Each prompt carries five-part anatomy, variants, parent-child composition, evals, and ownership so teams ship consistent output that survives model updates and new hires.
Prompt Pack
A prompt pack is the model-facing brand system that bundles system instructions, token references, few-shot examples, negative constraints, and output rules so AI can generate assets without inventing its own identity on every call.
Prompt Pattern
A prompt pattern is a four-part formula that declares format, purpose, audience, and constraints so AI tools like Figma Weave output structured layout scaffolds instead of generic wireframes.
Prompt Pinning
Prompt pinning is the practice of injecting fixed design system rules, tokens, and constraints at the top of every prompt or through persistent tool features so AI coding tools produce consistent output instead of drifting styles.
Prompt References
Prompt references are the specific artists, brands, products, eras, and files you name in a prompt to anchor the AI in exact taste instead of letting it default to training data averages.
Prompt Role
The opening instruction that assigns the AI a specific senior professional identity with named experience and clients. It stops generic slop and forces the model to operate with calibrated taste from the first token.
Prompt Sensitivity
Prompt sensitivity is the hair-trigger way Google Stitch and 2026 generative UI tools turn tiny wording changes into completely different layouts, layer structures, and component choices.
Prompt Surface
The full UI component surrounding an AI text input with empty states, suggestions, attachments, model pickers, tool toggles, streaming output, and revision controls that turns prompting into a structured, observable interaction.
Prompt System
A versioned library of reusable prompts with built-in evals that map to real design workflows. Prompt systems treat prompts like code or components and give orchestrators consistent measurable output instead of one-off guesses.
Prompt To Product
Prompt to product is the 2026 workflow where a plain English description of UI and behavior turns into production React code, wired logic, and a deployable app in minutes using v0, Bolt, Lovable, Cursor, and Claude Code.
Prompt Variant
A prompt variant is a configured version of a core prompt spine that adapts size, state, or role while sharing the same system prompt, examples, constraints, and output schema. Designers multiply one tested asset into many tools the same way Figma variants turn one button into nine usable states.
Prompt Versioning
Prompt versioning treats prompts as versioned assets using semantic versioning in git with mandatory evals on every change so teams know exactly why output quality shifts when models or rules update.
Prompt-to-UI
The practice of feeding a text description into an AI model to generate a responsive UI layout with real Figma layers instead of flat images or raw code.
Proof of Work
Proof of work shows your product performing live on the landing page instead of listing features in boring paragraphs. It replaces claims with demonstrations that force visitors to confront the reality of your tool.
Proof Strip
The proof strip is the band of logos or hard stats placed directly under the hero that delivers instant credibility without forcing the visitor to read a single word.
Proportional Harmony
Proportional harmony is the shared x-height, cap-height, and stroke weight that lets two structurally different fonts sit together without one looming or shrinking. It turns contrast into cohesion so pairings survive real layouts instead of fighting on the baseline.
Prototype To Production
Prototype to production is the workflow of converting a high-fidelity Figma prototype into functional code, typically a working Next.js page, in a single day using AI tools and code reading skills.
Proximity
Proximity groups interface elements by closeness. Items near each other read as related while items spaced apart read as separate categories. The whitespace does the organizing before labels or borders ever enter the picture.
Public Trail
A public trail is the living network of one shipped product, short decision logs, public writing, GitHub commits, annotated Are.na saves, and weekly social posts that prove taste, judgment, and shipping velocity in 2026.
R
Range Setter
The second project on your portfolio homepage that visually proves you can shift design languages without breaking a sweat.
Rate Bands
Rate bands are the real 2026 price ranges for logo work, full brand identity, and complete brand systems broken down by five studio tiers from new freelancers to Pentagram-level operators.
Rationale Overload
Rationale Overload is the mistake of overwhelming clients with every step of your design process instead of tying final choices directly to the approved brief. It invites nitpicking and scope creep by giving stakeholders too many loose threads to pull.
Raw Tokens
Raw tokens are the pure numerical values generated by your base size and ratio. They sit at the bottom of the token stack as the single source of truth that semantic and component tokens reference so one edit updates every platform without hunting through files.
Read Path
The exact sequence the human eye follows across a web page, independent of DOM order or grid position.
Reader Respect
The principle of treating your audience as intelligent adults who can handle complexity without progress bars, TLDRs, or other condescending training wheels.
Reading Order
Reading order is the specific sequence your visual hierarchy forces the eye to follow on any layout. It uses size, contrast, position, color, and spacing to make sure users see the hero message first, supporting details second, and your CTA last instead of inventing their own random path.
Reasoning Surface
A reasoning surface exposes an AI model's plan, steps, and chain of thought in plain language while it works. It turns opaque latency into readable logic users can scan, trust, and sometimes steer.
Reasoning Tokens
Reasoning tokens are the invisible, internal thought processes an AI model performs, billed at the standard output rate but often not displayed to the user. They represent the depth of the model's cognitive effort and problem-solving, not the length of its visible response.
Rebrand
A rebrand rewrites the strategic foundation including positioning name and visual system so the brand means something genuinely new.
Rebrand Backlash
Rebrand backlash is the market revolt that erupts when a new identity lands too far from the image already cemented in people's heads. It exposes the exact distance between your controlled design system and the messy collective perception you cannot simply redesign away.
Rebranding
The process of changing a brand's visual identity, messaging, positioning, or all three to better reflect what the brand has become or wants to become.
Recognition Management
Recognition management is the deliberate practice of mapping which visual and verbal cues actually trigger instant brand recall then protecting or phasing those cues during a refresh or rebrand so you do not throw away equity you already paid for.
Recoverability Affordance
Recoverability affordance is the set of in-place controls that let users edit, regenerate, undo, or save generative UI output at the level of individual intent slots and primitives.
Redesign Comms
Redesign comms is the narrative, timing, assets, and audience segmentation that explain why you changed the logo before customers encounter the new mark and start guessing.
Redesign Depth
Redesign depth is the strategic choice between a logo refresh, evolve, or full rebrand based on the specific business problem the current mark fails to solve.
Reduction Process
The systematic practice of removing non-essential elements from a design until only the elements that carry meaning remain. The core method behind minimalist logo design.
Reduction Test
The reduction test cuts every element that is not earning its place until the design breaks then restores the smallest piece that revives it. This protocol from Rick Rubin and Dieter Rams exposes the floor of your work and builds the judgment AI cannot replicate.
Redundant Entry
WCAG 2.2 AA criterion 3.3.7 that prohibits asking users to enter the same information more than once in the same transaction by requiring prefill, review steps, or selection from prior data.
Rejection Rate
Rejection rate is the ratio of AI-generated design variants you kill with written reasons to the ones you ship. In 2026 it became the clearest signal of taste because every designer generates polished work yet only those with judgment know exactly what to throw away.
Research Note
A dated, owned record that captures what was tried in a design or model experiment, what the results showed, what got rejected, and what the next move should be.
Research Synthesis
Research synthesis turns raw user interviews, tagged tickets, support logs, and past studies into clustered themes, insight statements, and roadmap bets. It used to eat two weeks of senior designer time at Meta and Airbnb. AI now delivers L5 output in four hours with Linear summaries, Notion AI, and targeted Claude prompts.
Responsive Collapse
Responsive collapse is the controlled reorganization of bento grid cells from desktop ratios through tablet two-column to mobile single-column stack that keeps the anchor first and the hierarchy intact.
Responsive Grid
A responsive grid is one grid system that collapses intelligently from 12 columns on desktop to 8 on tablet to 4 on mobile while preserving column alignments baseline rhythm and overall layout logic across every breakpoint.
Responsive Hierarchy
Responsive hierarchy adapts the five levers of visual hierarchy across every breakpoint so the eye follows the intended order on desktop, tablet, or mobile instead of flattening into noise.
Responsive Logo
A responsive logo is a system of purpose-built marks engineered for every scale from 16px favicons to hero headers instead of one master file that gets crushed or bloated by context.
Responsive Reflow
Responsive Reflow is the intentional collapse and reordering of bento grids from multi-column spans on desktop into a single-column stack on mobile that preserves hierarchy by surfacing hero cells first according to content priority rather than DOM position.
Responsive Type Behavior
Responsive type behavior defines how your text dynamically adapts its size, spacing, and presentation to maintain legibility, hierarchy, and brand consistency across all screen sizes and devices.
Retainer Pricing
Retainer pricing is a monthly fee for ongoing access to a design partner's capacity, AI systems, and compounding deliverables like prompt packs and token libraries. In 2026 it turns one-time AI infrastructure into predictable revenue that gets more valuable every month instead of burning out on project roulette.
Reveal Animation
The simplest and most universal logo animation pattern. A reveal uses scale, opacity, and custom easing to deliver the mark cleanly to a perfect still frame without drawing attention to the mechanics.
Reveal Moment
The reveal moment is the deliberate pause built into packaging between opening the outer layers and first seeing the product. It uses tissue, printed lids, pull tabs, or nested trays to control discovery rhythm and turn buyers into content creators.
Reversibility
Reversibility gives users one-click undos, regenerates, diffs, and preview gates so every AI mistake stays cheap to fix. It turns every output into a proposal and every action into a contract the user can reject without fallout.
Review Gate
A review gate is the structured checkpoint between workflow stages that names one decision, one human owner, one artifact and one criterion so taste survives acceleration.
Review Skip Failure
Review skip failure is when polished AI output tricks designers into approving work without enforcing the four-pillar review gate. The comp looks done so taste responsibility evaporates and regressions ship.
Role Naming
Role naming labels each type size by its job in the UI such as body, h3, or display instead of pixels or t-shirts. This creates a stable contract that survives changes to your base size or ratio.
Role-Based Tokens
Role-based tokens assign colors by UI purpose rather than hue or step number. Tokens like text-contrast, on-primary, bg-subtle and border-focus carry baked-in contrast ratios that satisfy WCAG AA the moment you use them.
Rollout Plan
A rollout plan catalogs every touchpoint where the old logo lives, assigns owners and hard deadlines to each, then sequences the swap from old to new so the brand stays coherent instead of fracturing in public.
Routing Layer
The decision service that reads every request, checks task type, context size, stakes, and budget, then fires it at the specialist model that actually wins on cost and quality instead of forcing one flagship to do everything poorly.
Rubric-Based Eval
Rubric-based eval scores AI outputs against a fixed list of explicit criteria instead of vague gut checks or single-score benchmarks. Claude 4.7 finally made it reliable enough to run as production regression tests on every prompt change.
Running Surface
The live codebase and running app where designers edit real components instead of static Figma frames. It delivers production fonts, actual breakpoints, and forced edge cases because the artifact you design is the artifact users touch.
S
SaaS Monoculture
SaaS Monoculture is the viral spread of the Linear aesthetic across B2B software in 2026 turning distinct products from Vercel, Resend, Cal.com, Cursor, and entire YC batches into visual clones through dark interfaces, gradient heroes, monospace accents, command palettes, and geometric grain. It is the homogenization loop where AI tools trained on the same fifty homepages make differentiation harder every cycle.
Safe Zone
Safe zone is the enforced clear space around a logo that travels with the mark on every surface. It is the operational version of active negative space that stops the brand from turning into noise the moment it leaves the brand book.
Safeguard Tax
The performance and transparency penalty designers pay when Anthropic's classifiers on Claude Fable 5 trigger a refusal and silently downgrade to Opus 4.8 mid-task without clear alerts.
Sales Gate
A sales gate replaces a clear price with a Contact Sales button on a pricing tier. It forces the buyer into a call before they can self-qualify and signals that the team dodged its packaging homework.
Sans Serif
Sans serifs are typefaces without the small finishing strokes on letter terminals. They rely on even weights, open counters, and precise spacing to deliver rational clarity for interfaces, dense data, and brands that signal engineering precision.
Saved Views
Saved views are persistent, URL-addressable bundles of filters, sorts, columns, and layout settings that let experts reconstruct their exact workspace instantly across devices and teammates.
Scaffold Quality
Scaffold quality is the functional but unpolished code AI app builders generate in minutes. It delivers a working structure with real auth databases and Stripe flows yet carries the fingerprints of a rushed first pass that demands heavy rewrites before it sees real users or real money.
Scalability
A design's ability to maintain clarity, impact, and legibility across all reproduction sizes, from a 16px favicon to a highway billboard.
Scale Collapse
Scale collapse happens when a detailed logomark looks strong at large sizes but disintegrates into noise at favicon and app icon scales. Too many elements, fine strokes, and competing shapes make it unusable exactly where brands live.
Scale Compression
Scale compression tightens your type scale ratio on mobile viewports to stop giant desktop headings from devouring limited screen space while preserving the original hierarchy.
Scale Ratio
Scale ratio is the single multiplier locked in at the start of any typographic system that generates every text size from one base value so contrast becomes predictable instead of improvised.
Scope Chip
A small UI pill that declares the exact scope of an AI memory entry or session so users always know whether something applies to this chat, this project, or everything they have ever told the agent.
Scope Creep
Scope creep is the uncontrolled addition of deliverables after the contract is signed and the price is locked without any extra budget or timeline. In brand identity work it turns a cleanly scoped 60,000 dollar project into 95,000 dollars of unpaid labor disguised as client service.
Scope Ladder
A Scope Ladder locks in the exact request types, active request limit, and turnaround commitment for a productized design service. It replaces ambiguous retainers with clear rules that protect capacity and enable predictable scaling.
Scoped Prompt
A prompt constrained by a scope token or selection that precisely limits what the model can see or edit without forcing the user to explain it in words.
Screenshot Friendliness
The metric that rewarded scroll-stopping mockups over products that actually worked. Designers optimized for what looked good in a Twitter screenshot instead of retention, trust, or long term value.
Screenshot Library
A screenshot library is a structured collection of captioned product screenshots from companies like Linear, Vercel, Arc, Stripe, and Notion that replaces moodboards as the primary reference system for AI powered design. Each entry follows strict naming conventions, includes your one sentence insight on the exact pattern worth stealing, and sits in shallow categories so models can query it for synthesis, token extraction, and recombination.
Screenshot Spec
A product specification built around real screenshots from shipping interfaces that serve as the primary source of truth, combining the image, a human-written caption, AI-extracted tokens, and strict constraints so engineering ships exactly what you meant.
Screenshot-Driven Design
The workflow that starts every design task with a real product screenshot fed directly into an AI model to extract patterns, tokens, and structure instead of beginning with wireframes or moodboards.
Screenshot-to-Code
Screenshot-to-Code is the process of generating functional, editable user interface code directly from a static image of a design. Recent advancements, particularly with models like Claude Fable 5, have elevated this from a crude approximation to a system-aware, high-fidelity translation.
Screenshottable Surfaces
Screenshottable surfaces are product interfaces designed from the first wireframe to read as complete branded ads when cropped into a social feed. They maintain clarity at thumbnail scale through ruthless visual hierarchy, one resilient focal point, and brand cues integrated so deeply the source is obvious without a caption.
Scroll-pinned Reveal
A scroll-pinned reveal locks a product interface inside the viewport while the rest of the page scrolls past it, updating the pinned panel in real time to demonstrate different states or features as the user progresses.
Scrolljacking
Scrolljacking hijacks natural scrolling to force users through designer-controlled animations and narratives, turning the scroll wheel into a video playhead instead of a content navigator.
Search-Empty
Search-empty is the screen that appears when a user runs a query and gets zero matches. It carries real user intent in the form of the exact search term, giving you context no other empty state provides.
Secondary Archetype
A secondary archetype supplies supporting texture to your primary archetype. It adds range and humanity while the primary promise stays ruthlessly in charge and every decision runs through it first.
Secondary Lockup
The alternate lockup for surfaces the primary does not fit, usually by switching from horizontal to stacked orientation or simplifying details so the brand stays intact instead of getting stretched or cropped in production.
Security Surface
The total attack area of a vibe-coded product including exposed keys, unprotected routes, missing rate limits, and unmaintained dependencies that AI generation rarely addresses.
Selection Driven Prompt
A selection driven prompt lets the user pick an object first then opens a prompt surface with that object already bound as live context so the model operates on the exact target without any manual description.
Selector Stability
Selector stability ensures agents can reliably locate and activate elements in your UI across updates. It demands semantic HTML, persistent data-testid attributes, and unchanging accessible names so the machine plane stays dependable even when the human plane gets redesigned.
Semantic Colors
Semantic colors map specific hues to explicit meanings. Success gets its green family, warning its orange, critical its red, info its blue. They are built as complete sibling systems after the neutral scale and anchor lock in so every variant communicates status instead of decorating the screen.
Semantic Layer
The semantic layer is the meaning-making middle tier in any real token architecture. It turns raw primitives into purposeful roles like text-primary and surface-elevated while hosting every light dark and brand mode in one collection.
Semantic Token
A semantic token assigns a specific role like surface-default or interactive-primary to a primitive value so components reference intent instead of raw hex codes or pixels.
Semantic Tokens
Design tokens that assign meaning to raw values. Instead of referencing color-blue-500 directly, components reference color-primary, which resolves to the appropriate raw value.
Sensible Defaults
Sensible defaults are the product's opening moves. They are the team's explicit bet on the right configuration for 80 percent of users so the other 95 percent never touch settings.
Sequential Color
Sequential color uses one hue ramp from light to dark or low saturation to high saturation to encode ordered magnitude from low to high.
Serif
A serif is the small finishing stroke at the end of a letter's main strokes. It separates typefaces that carry four centuries of heritage signaling from those that read as clean, rational, and built for systems.
Session Reset
Session reset is the deliberate act of abandoning a bloated AI chat and starting fresh with only the distilled essentials pulled into a clean document. It treats the conversation as temporary RAM instead of permanent storage so quality stays high and costs do not spiral.
Settings Dump
A settings dump piles toggles chronologically with no hierarchy or grouping as a fossil record of every decision the team avoided.
Settings Labyrinth
A settings labyrinth buries toggles under endless nested tabs, subcategories, advanced drawers, and modals. Users know the setting exists because they found it once but still cannot locate it without wasting ten minutes because the team refused every hard call on defaults, priority, and hierarchy.
Settings Museum
The Settings Museum is the graveyard wing of any mature product where toggles for long-dead features sit untouched for years because nobody wants to own the deletion.
Settings Sprawl
Settings sprawl turns team indecision into endless toggles, buried menus, and abandoned code paths. The settings screen becomes a landfill of fossils that reveals every argument the team avoided.
Shadcn
Shadcn is a copy-paste collection of React components built on Radix primitives and styled with Tailwind that lives directly in your repo instead of as an npm dependency.
Shadow Stacks
Shadow stacks layer a contact shadow, ambient shadow, and directional shadow to mimic how real objects sit under light and cast believable depth.
Share-Multiplier Outputs
Share-multiplier outputs turn one user task into multiple ready-to-share artifacts. The product auto-generates tweet cards, branded screenshots, interactive embeds, and rich previews at completion so users broadcast their wins and pull strangers back to the source without extra effort.
Shared Element Transition
A shared element transition animates the exact same UI component from one layout to another so the user watches it travel, scale, and settle instead of seeing it disappear and reappear. The card you tap becomes the detail header in one continuous move, teaching spatial relationships without extra copy.
Shelf Impact
The ability of packaging to attract attention and communicate brand identity from typical retail viewing distance (approximately three feet). The primary performance metric for retail packaging.
Ship Date Ownership
Ship date ownership is the designer naming a hard calendar date for launch and carrying full accountability for hitting it. The 2026 designer-PM writes the date into their spec, makes every tradeoff to protect it, coordinates directly with engineers, and owns the outcome instead of waiting for a PM to set it.
Shipped Product
A live digital tool with real users, a working URL, and visible usage data that proves a designer can make tasteful calls under actual constraints.
Shipped Work
Shipped work is design that survived engineering, QA, launch, and real users. It comes with production screenshots, live URLs, and hard metrics instead of Figma comps and good intentions.
Shipping Senior
The Shipping Senior is the designer who reads the codebase, opens PRs, and ships deployed product end to end. Figma is a fast sketch. The deliverable is production code.
Silent Failure
Silent Failure is when an MCP server crashes or never starts yet Claude Desktop or Cursor shows zero errors. The tool panel stays empty, commands return nothing, and the AI works from memory while pretending the connection succeeded.
Silhouette
The solid outline shape of a design when filled with a single flat color. A strong silhouette means the mark is recognizable without internal detail, color, or texture.
Similarity
Similarity is the Gestalt principle where elements that share visual characteristics get grouped by the brain before conscious thought kicks in. Same look means same job.
Simplified Mark
The simplified mark is the third tier in a responsive logo ladder, a complete redraw of the primary mark stripped of every detail that fails below 80 pixels so it remains crisp and recognizable in UI components and thumbnails.
Simplified Wordmark
The simplified wordmark pairs your logomark with a condensed, abbreviated, or redrawn version of the brand name built specifically for medium and small spaces where the full wordmark collapses into noise.
Simultaneous Contrast
Simultaneous contrast is the optical effect where adjacent colors alter each other's perceived hue, brightness, or saturation. The human visual system exaggerates differences to create sharper boundaries which is why the exact same swatch can look vibrant on one background and dead on another.
Single Decider
The one named person whose approval actually ends the design project. No committees. No founder feedback loops at the end. One human owns the yes.
Single Family System
A single family system uses one rich typeface family and its full range of weights widths and styles to handle every role from headlines to body copy to UI instead of mixing multiple different typefaces.
Single Revelation
Single revelation is the principle that negative space logos work best when they contain exactly one hidden element. Multiple revelations overload the mark turn it from memorable to merely clever and destroy the three-second recognition window logos actually get.
Single-Frame Output
Single-frame output is what every prompt-to-UI tool in 2026 still defaults to: one isolated responsive screen per prompt instead of connected flows, interactive states, or multi-screen prototypes.
Single-Screen Today View
The home screen that answers exactly one question: what should I handle right now. It collapses identity, time, and context into a single scroll-free surface of next actions instead of dumping every metric in a neutral grid.
Size Lever
Size lever makes the most important element on the page the largest by a stupidly obvious margin so the eye lands there first every single time.
Size Threshold
Size thresholds are the exact pixel breakpoints where one logo tier hands off to the next simpler one in a responsive system. They stop your detailed wordmark from collapsing into noise at 16px and keep chunky icons from looking clumsy when scaled large.
Size Trigger
The exact pixel breakpoint where one tier of a responsive logo system swaps for the next to preserve legibility and brand clarity.
Skeleton Screen
A placeholder layout that mirrors the exact shape of final content using gray blocks and lines so users orient themselves while data loads.
Skeuomorphism
Skeuomorphism is the design approach that copies real world material physics so pixels feel like physical objects users already know how to use.
Skill Description
The YAML frontmatter line in SKILL.md that tells Claude exactly when to load the Skill and what job it performs.
Skill Evaluation
Skill evaluation is the structured test process that runs a Claude Skill against real past projects with known ground truth outputs to verify it catches every relevant issue, ignores what it should ignore, and invents zero false positives before it ships.
Skill Library
A skill library is a versioned collection of Claude Skills stored in a Git repo like a design system for prompts. Each Skill is a scoped folder with a SKILL.md trigger file and reference docs so Claude loads the latest rules automatically instead of forcing the team to retype stale prompts.
Skill Pack
A bundled set of Claude custom instructions, MCP tool definitions, taste profiles, critique rubrics, and real client examples that clones a senior designer's exact process so it can run at scale or sell as a standalone product.
Skill Scoping
Skill scoping limits each Claude Skill to a single tightly defined task so the model loads precise instructions at the right moment without bloating context or drifting in quality.
Skip Link
A hidden link at the top of the page that lets keyboard users jump straight to main content instead of tabbing through every header and navigation item.
SKU Color System
A systematic approach to differentiating product variants within a packaging line through color while maintaining overall brand recognition.
Slab Serif
Slab serifs are typefaces with heavy blocky serifs that match the weight of the main strokes. They deliver mechanical boldness and texture that bridges traditional serif structure with sans-like attitude making them punchy for display work.
Slant Axis
The slant axis (slnt) tilts letterforms across a continuous range of oblique angles from a single variable font file instead of forcing separate static cuts for every weight and style.
Small Multiples
Small multiples are a grid of identical charts repeated for different data slices so the eye can compare patterns across categories without overlap or chaos.
Smart Defaults
Smart defaults pre-fill most of the required information from a single signal like company name, email domain, or imported framework so the user starts with 90 percent of the work already done.
Social Avatar
A social avatar is the purpose-built circular brand mark that survives the hard round crop enforced by Instagram, X, LinkedIn, Threads, and TikTok. It centers a monogram or simplified symbol with exact padding so the shape reads in under a second at forty pixels instead of turning into a clipped mess.
Social Proof
Social proof is the strategic placement of logos, testimonials, metrics and product evidence that reassures visitors others have already succeeded with your offer.
Social-First
Social-first means you design vertical and square formats like 9:16 and 4:5 as the primary hero assets then adapt them upward to websites and banners. It treats the phone feed as the source of truth because that is where 80 percent of impressions actually happen.
Soft Degradation
Soft degradation is the gradual decline in AI output quality during long sessions that occurs well before any hard context limit. The model keeps answering yet responses grow repetitive, forget constraints, and lose sharpness without throwing errors.
Software Long Tail
The software long tail holds every niche workflow too specific for startups yet too painful for spreadsheets. Personal software finally populates it with tiny perfect tools built for named users instead of abstract markets.
Solo Ceiling
The solo ceiling is the hard 80 percent line where a designer using vibe coding tools can ship a real clickable product before senior engineering is required for security hardening, scalable architecture, payments infrastructure, and compliance.
Solo Design Studio
A one-person design practice that clears seven figures annually by stacking retainers, productized services, brand audits, and AI workflow consulting while using tools like Claude and Cursor to replace the junior production layer.
Solution Space
The width of possible valid answers to a given design problem, from narrow tasks with one right answer to wide open problems with many equally viable approaches and real tradeoffs.
Sonic Anthem
The 30-to-60-second brand theme that serves as source code for your entire sonic system. Expand your motif into a full emotional arc here, then extract the sting, derive the ambient bed, and set the tone for every UI sound.
Sonic Brand Identity
Sonic brand identity is the coordinated four-asset audio system that owns every sound touchpoint your logo cannot reach. Built from one motif and one brief it delivers instant recognition and emotional consistency where visuals never show up.
Sonic Brief
The five-section document that locks voice, use cases, tier, formats, and deliverables before a composer starts so the four-asset sonic system arrives coherent in two rounds instead of six.
Sonic Logo
A sonic logo is a one-to-three-second audio brand mark, the ear's version of a visual logo. It compresses a whole identity into something you recognize before a single pixel loads, which is why brands from Netflix to Mastercard build entire sound systems around one short mnemonic.
Space Lever
The space lever ranks elements by isolation. The element with the most negative space around it becomes the primary read regardless of its size weight or color.
Spacing Guess
Spacing Guess is the inconsistent application of line height, letter spacing, and paragraph spacing, leading to arbitrary visual rhythm and a disjointed user experience.
Spacing Standards
Spacing standards are the codified rules for line height, letter spacing, and paragraph spacing within a typography system, ensuring consistent legibility, visual hierarchy, and vertical rhythm across all contexts.
Spatial Reference
A spatial reference is an AI-generated layout scaffold from tools like Figma Weave that you pin to the canvas as a fixed guide for structure, rhythm, and hierarchy while building real production screens with your component library alongside it.
Spatial UI
Spatial UI deploys precise 3D moments on the web to add tangible depth that clarifies form and function. It favors small targeted interactions like scroll-synced rotations and hover tilts over heavy hero scenes.
Spec Anatomy
Spec anatomy is the consistent seven-section structure used in high-performing design specs: intent, scope, behavior, edge cases, success criteria, evals, and accessibility plus copy.
Spec Scope Seed Ship
Spec Scope Seed Ship is a four part prompt pattern that produces shippable code by first declaring the exact technical context, then setting hard boundaries, anchoring to visual references, and finally listing measurable acceptance criteria.
Spec-Driven Design
Spec-driven design is the practice of treating a tightly written text specification as the primary design artifact, with visuals, code, and tests flowing downstream from clear intent, behavior, and success metrics.
Spec-First
Spec-first is the workflow that starts with a complete written specification before any visual design or code, allowing AI tools and engineers to build from intent instead of iterating on mockups.
Specular Highlights
Specular highlights are the crisp 1 to 2 pixel white or near white lines placed along the top edge of elevated UI surfaces to signal real volume and a consistent light source.
Split Complementary
Split complementary pairs one dominant hue with the two colors that sit on either side of its direct complement on the color wheel. It delivers strong contrast with less visual violence than a pure complementary pairing, which is why it appears in so many mature brand systems that actually ship.
Split Screen Hero
The B2B hero pattern that puts a crisp narrative on the left and a self-evident product screenshot on the right. The interface becomes the proof.
Spring Physics
Spring physics uses physically modeled springs instead of timed curves to give UI elements realistic weight, overshoot, and settling that makes interfaces feel fast, precise, and alive.
Square Modular Grid
A square modular grid is a tile field of equal squares that forces every dimension of a logo to multiples of a base unit usually 8 pixels for unbreakable consistency in geometric marks.
Squint Test
A quick hierarchy check where you blur your vision (or squint) to see which elements stand out when detail is removed. If the right things pop, the hierarchy works.
sRGB
sRGB is the 1996 RGB color space that defines every hex code and rgb() value on the web. Built for CRT monitors by Microsoft and HP it maps RGB channels to light output using BT.709 primaries and a 2.2 gamma curve instead of matching human vision.
Stack Opinion
Stack opinion measures how rigidly an AI app builder forces its own technology preferences onto the code it generates. High stack opinion tools bake in their DNA whether it fits your project or not, turning the output into either a seamless extension of your stack or a permanent refactor battle.
Staging Environment
Staging environment is the final safe stop before code reaches real users. It runs production-like infrastructure with scrubbed realistic data so designers can catch every layout break, state failure, and interaction flaw that Figma never shows.
Stakeholder Map
A grid that plots every decision maker by power, motivations, fears, and brand references so you tailor the presentation to their exact incentives and kill objections before they kill your work.
Stale Context
Stale context is the accumulated noise of old directions, scrapped ideas, and outdated constraints left in an AI chat that distorts every new response. It forces the model to produce compromise outputs that split the difference between your current ask and yesterday's discarded plans.
Standalone Symbol
A standalone symbol is the distilled shape or letterform that carries your entire brand at sizes where words die, built from the ground up to dominate at 16-32px in favicons, app icons, and notification badges.
Stark
Stark is the Figma plugin that runs WCAG contrast checks, simulates eight types of color blindness, flags text size violations, and drops annotated fix layers straight onto your canvas.
State Colors
State colors are the hover, focus, pressed, and disabled variants paired to every interactive color in your system. They are built as predictable shifts on your anchor, semantic, or neutral scales rather than new hues invented at the last minute.
Static Font
A static font is a single-weight font file with fixed outlines and no axes. You need a separate file for every weight, width, or style you intend to use.
Static Frame
A single, non-interactive visual representation of a user interface, capturing one specific state, often used as input for AI code generation.
Static Site
A static site is a set of pages pre-rendered to plain HTML at build time and served straight from a CDN, so no server work happens when a visitor loads it. That makes it fast and cheap to host, with one tax: changing content means rebuilding the site, unless you bolt on incremental static regeneration.
Status Legibility
Status legibility makes every state change in your product both visible to humans and parseable by machines through structured events, named states, and ARIA live regions that sit alongside the visual layer.
Sticky CTA
A sticky CTA is a minimal persistent button or bar that follows the user after the hero scrolls out of view keeping the primary conversion action one tap away on long landing pages.
Still Frame
The non negotiable resting state every animated logo must land on cleanly. It survives when motion fails and serves as the anchor for all four motion modes.
Strategy Interview
The structured conversations with founders, marketers, and customers that convert vague logo complaints into a ranked list of what the mark must communicate.
Streaming Interrupt
The visible stop button, live accept/reject controls, and branching mechanisms that let users halt or redirect an AI stream the moment it goes wrong. It sits as layer four in every serious streaming UI and prevents users from enduring full generations of garbage.
Streaming Partial
Streaming partial renders the first usable chunk of server output the moment it arrives while the rest streams in behind it. Users start reading or acting on real content immediately so the loading state vanishes into active consumption.
Streaming Surface Audit
A seven point pre ship checklist that verifies first feedback, cadence, structural commitment, stop affordance, edit options, completion design, and post stream interactivity.
Streaming UI
The complete output surface that delivers AI content with rhythm, structural stability, interrupt controls, cursor behavior, and post-generation handoff instead of a raw token dump into a div.
Structural Commitment
Structural commitment locks in layout for lists, tables, code blocks, and other elements the instant they become detectable in a token stream so the page never reflows and the user's eye never hunts for its place.
Structural Contrast
The deliberate difference in typeface construction, pairing geometric with humanist or serif with sans, that creates visual interest without breaking harmony.
Structural Draft
A structural draft is an AI-generated UI layout skeleton from a tool like Google Stitch that delivers sound hierarchy, grid logic, named layers, and auto-layout in Figma so you start with real bones instead of a blank frame or a useless PNG.
Structured Output
Structured output is the AI pattern where the model returns a typed data object that your interface renders directly instead of a paragraph of text the user must parse and copy from a chat.
Structured Rubric
A tight set of five to seven measurable criteria scored one to five with mandatory one-line reasons, a hard pass threshold, and JSON output so an LLM can judge thousands of AI-generated candidates in minutes.
Style Dictionary
Style Dictionary is Amazon's open source build tool that takes one JSON file of design tokens and outputs perfectly formatted CSS variables, Swift files, Android XML, and more.
Sub-brand
A sub-brand earns its own name, personality, and visual rules while locking up with the master brand to borrow trust and equity instead of building it from zero.
Subagent
A subagent is an autonomous AI entity delegated a self-contained task, operating within its own isolated context window to deliver a specific result back to a main agent or user. It is a specialized worker designed to manage complexity and optimize token usage by handling heavy lifting separately.
Subheadline
A subheadline sits under the hero headline and specifies exactly who the page is for while naming the concrete outcome or change the right visitor will get. It turns the headline claim into something personal and believable before trust elements and CTAs take over.
Success Criteria
Success criteria are the numeric targets in a design spec that prove whether a feature actually solved the user problem instead of just looking better.
Success State
Success State is the concrete preview of a finished AI output that users see before creating their own. It removes guesswork by showing exactly what quality, structure, and depth look like so users can recognize success instead of staring at a blank bar wondering if their prompt even worked.
Suggestive Name
A suggestive name hints at a benefit or quality without stating it directly. It sits between descriptive and invented options to deliver instant stickiness, trademark strength, and room to grow.
Supervised Execution
Supervised execution is real-time human oversight of AI agents doing computer use tasks. The person watches every screenshot and action and can pause, correct, or seize control before mistakes destroy data or waste hours.
Surface Copy
Surface Copy lifts another product's interface pixels while leaving its decision stack untouched. The result looks like the original on a pitch deck and fights like the legacy tool it replaced the moment real work begins.
Surface Elevation
Surface elevation creates visual hierarchy in dark mode by lightening each raised UI surface rather than casting shadows. The technique replaces the light mode shadow system with a stepped scale of surfaces that step toward the viewer.
Surface Elevation Model
A surface elevation model defines four distinct lightness tiers for backgrounds, surfaces, elevated cards, and overlays in dark mode where traditional shadows fail to create depth. Each step uses small deliberate increases in lightness and controlled drops in chroma.
Surface Token
A surface token defines the background color and elevation for every container in your UI so the same component adapts to light mode, dark mode, or new themes without manual fixes.
Surrounding Margin
Surrounding margin is the outer whitespace that frames the complete bento grid so the differently sized cells read as one deliberate ranked unit instead of scattered blocks on the page.
Survivorship Bias Funnel
The survivorship bias funnel is the platform mechanism that elevated a tiny group of daily posters and thread gurus into perceived design authorities while erasing the much larger group of practitioners who shipped real work without ever trending.
SVG
SVG is the web-native vector format that keeps logos crisp at any resolution with zero extra files.
Symbol Stroke Weight
Symbol stroke weight is the calculated thickness of every line and edge in a logo symbol engineered to stay legible at 16px. It is a technical constraint set at minimum render size first so the icon tier survives favicons, notifications, and app grids without turning into generic mush.
Symmetry and Order
Symmetry and order is the Gestalt principle that makes balanced regular layouts feel resolved and competent before users read a word. The brain rewards grids that align and rhythms that repeat with instant trust.
System First
Design tokens, primitives, and patterns before you compose a single page. It stops the same interface decisions from being reinvented on every surface and turns shipping from constant reinvention into simple composition.
System Instructions
Hidden directives loaded into an AI session that shape the model's behavior, tone, and constraints. They consume context window tokens like any other input.
System Ownership
System ownership means holding the living design system in code as the single source of truth that both humans and AI query at companies like Linear, Vercel, and Anthropic. It turns taste into enforceable rules, tokens, components, and decision logs so output scales without linear headcount or quality collapse.
System Rules
The enforceable constraints that stop your brand identity from turning into generic slop six months after launch.
T
Tailwind v4
Tailwind v4 is the utility CSS engine that reads your CSS variables natively so tokens from Figma Variables flow straight into production classes without ever touching a config mirror.
Target Size
The minimum area for any tappable or clickable element so users can activate it without error. WCAG 2.2 requires 24 by 24 CSS pixels at AA level with exceptions for spacing.
Task Framing
Task framing is the structured input pattern that converts a user's casual goal into a complete set of editable parameters before an agent begins planning. It forms the foundation of trustworthy agent UIs by eliminating assumptions that lead to misaligned execution.
Taste Range
Taste range is the demonstrated ability to apply consistent design judgment across product interfaces, brand systems, motion, writing, and code instead of repeating a single visual or problem solving mode.
Teaching Illustration
A visual in an empty state that previews the populated interface using your exact UI components and realistic data so users instantly understand the feature before they add anything.
Technical Warm Voice
Linear's writing style that delivers engineering-grade precision with genuine human warmth. It reads like a sharp builder explaining real decisions to peers without hype, jargon, or corporate padding.
Ten-Year Test
The ten-year test projects a candidate name forward to 2036 to see if it still reads neutral on a business card instead of dated, cute, or trapped in its birth-year trends.
Tetradic
A complex color scheme built from two complementary pairs, forming a rectangle on the color wheel, offering rich visual diversity but demanding careful management.
Theme Inversion
Theme inversion is the lazy shortcut of flipping light mode colors to create dark mode by turning backgrounds black and text white without redesigning surfaces or contrast.
Thinking Text Loop
A thinking text loop is the rotating set of vague status messages like Thinking... Generating... Almost there... that AI interfaces spin while the model works. It fakes progress with zero actual information and trains users to ignore every status signal the product ever ships.
Thread Guru Pattern
The thread guru pattern is a rigid Twitter format built around a provocative hook, a confident hot take, a numbered list of tips, and a call to action that drove follows while delivering only a thin layer of actual craft at the bottom.
Three Layer Feedback
Three Layer Feedback sorts every critique comment into observation of what is literally on screen, interpretation of what it means for users or the system, and suggestion of what to try next so teams fix real problems instead of trading vague opinions.
Three Layer Model
A framework splitting brand into identity, image, and reputation to diagnose which layer is broken and which fix to apply.
Three State Toggle
A theme switcher exposing three options: system, light, and dark that reads OS preferences via matchMedia, writes overrides to localStorage, and connects to your semantic token layer without flashing the wrong palette.
Three Tier Hierarchy
Three tier hierarchy ranks every interface element into primary, secondary, or tertiary importance then assigns size, contrast, color, spacing, and position to match. It is the practical framework that prevents every element from screaming at once and guides users in under three seconds.
Three-Foot Rule
The Three-Foot Rule demands your packaging communicate brand, product, and value using shape, color, and one dominant element from three feet away. Fail this test and you lose the shopper before they ever pick up your product.
Three-Second Rule
The three-second rule is the test that forces every product chart to reveal its core insight to a new user in three seconds or less. Fail the test and the chart gets redesigned or killed.
Tiered Routing
Tiered routing sends each task to the right Claude 4.7 model. Opus tackles hard reasoning, Sonnet runs the main loop, and Haiku handles high volume simple work. The pattern that makes production agents both smart and affordable.
Time to First Token
Time to first token is the critical window between prompt submission and the first honest feedback signal in a streaming UI. Under 200ms with a real indicator the interface feels instant. Beyond 800ms with a blank screen and users assume the product is broken.
Time To Insight
Time to insight measures the seconds between opening a product and reaching an actionable conclusion about what changed and what to do. Most 2015 dashboards round this number to forever while Linear-style today views, generative summaries, and embedded data crush it to single digits.
Timeline Trained Taste
Timeline trained taste is the narrow visual and craft sensibility designers developed by absorbing years of Design Twitter threads and screenshots rather than by studying shipped products or wrestling with real constraints. It mistakes engagement for quality, surface mimicry for deep craft, and loud posters for working designers.
Token Discipline
Token discipline is the rule that every visible value in a design file, color, spacing, type, radius, shadow, motion, must resolve to a semantic token from a single source of truth.
Token Drift
Token drift occurs when your Figma tokens and production code quietly part ways. One side updates a radius, shadow, or color value while the other stays frozen, creating visual inconsistencies that compound until the product feels incoherent.
Token Graph
A token graph turns every brand decision into connected machine-readable nodes so AI can pull exact values for color type voice motif grid and ratio without guessing.
Token Layer
The token layer is the semantic foundation where raw color primitives become role-based decisions that carry baked-in contrast ratios, perceptual rules, and allowed surface pairings so accessibility travels automatically with every UI choice.
Token Layers
Token layers are the three-level hierarchy that separates raw global values from semantic meanings and component-specific bindings so you can change one thing and update thousands without hunting down stray hex codes.
Token Naming Convention
A four-part system using category, role, variant, and state that creates readable self-documenting names for every design decision in your system.
Token Propagation
Token propagation is the disciplined process of updating one design token at its source and reliably updating every component, story, package, and site that depends on it without leaving hardcoded stragglers or introducing regressions.
Token Remapping
Token remapping is the immediate swap of every AI-generated token from tools like Google Stitch for your own design system tokens inside Figma so colors, typography, spacing and radii match your product instead of generic training data.
Token Reuse
The compounding effect where each new AI response requires reprocessing all previous conversation tokens, increasing latency and cost with every turn.
Token Rhythm
Token rhythm is the smoothed visual cadence of streaming text created by buffering raw model tokens and rendering them at a steady pace tuned to human reading speed with an anchored cursor for guidance.
Token Scoping
Figma's mechanism to control where variables can be applied so designers cannot accidentally bypass your architecture by grabbing primitives directly.
Token Snowball
Token snowball is the compounding growth in input costs as every new message in an AI session forces the model to reread the entire conversation history. What starts cheap quickly becomes expensive.
Token Sweep
Using Cursor's agent mode to intelligently update every instance of an old design token across hundreds of files in one pass, understanding semantic relationships instead of doing dumb string replacement.
Token Theming
Token theming swaps primitive values behind semantic token names so light dark and high-contrast modes all run on the exact same components.
Tokens Studio
Tokens Studio is the Figma plugin that exports Variables to version-controlled JSON, Style Dictionary, or Tailwind configs your engineers actually ship. It is the only tool that keeps design systems honest between Figma and production code at scale.
Tokens Upstream
Tokens upstream means the single source of truth for every design decision lives in a codebase file that both Figma and production code read from, so changing a value once updates every surface downstream without handoffs or drift.
Tone Context Map
A tone context map plots how your fixed brand voice adjusts across real situations like marketing, error states, onboarding, and support so writers and AI stop guessing.
Tone Spectrum
The plotted axes that lock in exactly how your brand should feel across every touchpoint so writers and designers never have to guess.
Tool Layering
Tool layering means assigning each AI coding tool to the exact layer of the stack where it dominates instead of forcing one tool to do every job. v0 owns UI fidelity, Bolt owns quick prototypes, Lovable owns founder MVPs with real data, Cursor owns real codebases, and Claude Code owns heavy refactors across dozens of files.
Tool Surfaces
Tool surfaces are the named actions agents invoke by verb. They turn every primary button, workflow, and capability into a stable contract where the human label, ARIA name, selector, API endpoint, and MCP handler all speak the exact same language.
Tool Use
Tool use is the AI pattern where models call structured functions with exact JSON parameters and receive clean data back. It skips screenshots and mouse clicks entirely, delivering ten times lower cost and five times faster execution than computer use whenever an API exists.
Top Navigation Bar
The horizontal header strip that holds primary links visible on every desktop page so users can orient and move without hunting.
Tracking
Tracking is the uniform spacing applied to every letter in a run of text. It tightens large headlines for density or loosens small uppercase labels for intent but must remain at zero for body copy.
Translation Tax
Translation tax is the accumulated time, fidelity loss, and organizational drag that happens every time a designer must explain visual intent to someone else for implementation.
Transparency Surface
Transparency surfaces display exactly what context the model used, which decisions it made, and the sources behind its outputs so users can verify accuracy without guesswork.
Trend Handcuff
A logo trapped in the dominant aesthetic of its launch year. It looks current for one season then dates the brand for the next five because it copied the Dribbble trend instead of solving the actual identity problem.
Triadic
A color scheme composed of three hues spaced equally around the color wheel, forming an equilateral triangle and delivering balanced, vibrant energy.
Triadic Harmony
Three hues spaced exactly 120 degrees apart on the color wheel that deliver vibrant balanced energy. One color leads while the other two support in muted roles.
Trust Pattern
A trust pattern is a UX surface that exposes AI reasoning, sources, limits, and reversibility so users can verify output in the first five minutes instead of gambling on hallucinations.
Trust Signals
Trust signals are review surfaces that show exactly what an agent changed so humans can approve or reject before anything commits. Cursor diffs, Claude artifacts, and Copilot plans turn scary autonomy into safe collaboration.
Twelve Column Grid
A twelve column grid divides any layout canvas into twelve equal vertical columns with fixed gutters so every element snaps to whole number spans that solve nearly every common web pattern without fractions.
Two Layer Token System
A two layer token system splits raw color values into primitives and semantic roles. Primitives store the actual OKLCH numbers while semantics act as stable aliases that components reference allowing you to swap entire palettes by remapping primitives alone.
Two Line Pitch
The two line pitch is the 15-second opener at the top of every case study that states the client, the problem, your exact role, and a tied outcome so a hiring manager knows the project's value without reading further.
Type Longevity
Type longevity is the half-life of a typeface's most distinctive details before they start dating the entire brand. It calculates how many years your display choice survives cultural shifts before it triggers an expensive refresh nobody budgeted.
Type Scale
A set of font sizes generated from a consistent mathematical ratio. Instead of picking sizes by feel, you pick a base size and a ratio, and every other size flows from that relationship.
Type System Problem
An inconsistency in a product or brand's typography that arises from a lack of defined rules for how text behaves across different contexts and applications.
Type to Confirm
Type to Confirm forces users to manually type the exact resource name or the word DELETE before an irreversible action activates, converting muscle memory into conscious intent at the highest-stakes moments.
Type-Driven Hero
A type-driven hero lets massive raw typography own the entire viewport in a brutalist layout, replacing imagery, illustrations, and UI chrome with letterforms that do the full branding job.
Typographic Ratio
A typographic ratio is the single multiplier you apply to a base font size to create every larger or smaller step in your scale so sizes feel perceptually balanced.
Typographic Register
Typographic register determines how loudly your brand speaks through type. It sets the amplitude from quiet authority to loud expression separate from the stylistic voice of the typeface.
Typographic Voice
Typographic voice is the immediate personality projected by letterform shapes, stroke contrast, terminals, and proportions before the words themselves are read.
Typography Audit
A typography audit is the five-question stress test that reveals whether your typeface actually signals the brand position you claim or quietly lies about it to every customer who lands on your site.
Typography Hierarchy
The system of font sizes, weights, and styles that guides the reader's eye through content in order of importance.
Typography Scale
A system of proportional text sizes that creates consistent visual order across headings, body text, captions, and labels. Usually based on a mathematical ratio.
Typography System
A typography system is the complete set of rules governing scale, font roles, weights, spacing, and responsive behavior so every piece of text stays consistent across every surface your brand touches.
U
UI Designer
The designer who turns validated decisions into coherent, interactive screens. They own visual systems, component states, motion, and the thousand micro details that make a product feel like one product.
UI Fidelity
UI fidelity is designing interfaces directly in the production codebase so every font, component, breakpoint, hover state, and edge case matches exactly what users experience instead of faking it with static Figma frames.
UI Primitives
Headless accessible building blocks from Radix and Shadcn that handle behavior, focus management, and keyboard navigation so you own the visuals and taste instead of reinventing the wheel.
UI Sound
UI sounds are the sub-second branded audio cues that respond to every tap, swipe, success, error, and state change in your product. They replace OS defaults with the same motif as your logo sting so daily use reinforces your identity instead of Apple's.
Ultracode
The highest setting in Claude Code that triggers multi-agent workflows where several instances plan, write, critique, and revise instead of delivering a single pass answer.
Ultrathink
A one-off prompt injection that forces a single deep reasoning pass in Claude without changing your persistent effort level for the whole session.
Unboxing Experience
The designed sequence of moments a customer encounters when opening a product package. A brand touchpoint that reaches audiences beyond the buyer through social media sharing.
Undo Window
An undo window is a timed grace period after an action executes during which the user can still reverse it. It lets products move fast on reversible tasks while protecting users from immediate regret without blocking them upfront.
Unicode Range
Unicode range is the @font-face descriptor that lists specific character code points so the browser downloads a font file only when those exact characters appear on the page.
Unlimited Design
A subscription where clients pay a fixed monthly fee for as much design work as they submit. The term sounds liberating until real queues expose its limits.
Up Stack
The strategic shift of moving design work from production tasks to higher-order editing, selection, and judgment as AI commoditizes making.
Usage Object
The `usage` object is a JSON payload returned by AI APIs, like Anthropic's Claude, detailing the exact token consumption for a given interaction. It breaks down input and output tokens, often including specifics like cache reads and writes, serving as the definitive record for billing.
Use Case Grid
The landing page section that maps your product to specific buyer scenarios so visitors instantly see themselves in it instead of guessing at fit.
User Research
User research is the practice of studying real users through interviews, usability sessions, surveys, analytics, and session replays so design decisions rest on evidence instead of the loudest opinion in the room. It splits into generative research that discovers what the problem actually is and evaluative research that tests whether your solution works.
UX Debt
The gap between flows that feel obvious to the solo founder who built them and flows that actually work for first-time users who lack the builder's context.
UX Designer
The designer who owns decision architecture. They decide what the product needs, what order it appears in, and what happens when users get confused, wrong, or distracted.
V
Value Clustering
Value clustering is the visual system's tendency to group elements of similar brightness together while separating them from elements with different brightness. It makes value contrast more powerful than hue contrast for establishing hierarchy.
Value Contrast
The difference in lightness or darkness between two or more colors, crucial for creating visual hierarchy, readability, and depth within any color scheme.
Value-Based Pricing
A pricing model that sets the fee as a function of the outcome the client gets, not the hours the work takes or the cost of producing it.
Values Driven Identity
Values Driven Identity lets core beliefs shape every visual, verbal, and strategic decision instead of trends or aesthetics. The conviction becomes the only brief needed.
Vanity Dashboard
Vanity dashboards pack screens with green charts and rising metrics so executives can screenshot them for decks instead of surfacing real problems that need fixing.
Variable Collections
Variable collections group Figma tokens by distinct concerns so modes can handle variations without duplicating structures or breaking alias chains.
Variable Drift
Variable drift occurs when Figma design tokens and their code implementations fall out of sync over time. MCP connected AI agents like Claude Code detect these mismatches instantly by reading real variable references instead of screenshots.
Variable font
A font file that exposes one or more axes (weight, width, slant, optical size, or custom) that can shift at runtime without loading additional files.
Variable Modes
Collections of alternative values inside a single Figma variable collection that let you swap entire themes like light to dark without duplicating your tokens.
Variant System
The full set of logo lockups, orientations, color versions, and usage rules that turn a single mark into something that works reliably across every surface from favicon to embroidery.
Vector Source
The vector source is the original editable master file, usually an .ai, that holds your logo as mathematical paths, layers, and live elements. Every SVG, PDF, PNG, and EPS is merely an export from this single source of truth.
Vendor Lock-in
Vendor lock-in is when a platform's proprietary guts turn speed today into a painful full rebuild tomorrow. Webflow creates the heaviest version in 2026 because its layouts, class names, and interactions do not export cleanly. Framer is lighter. Next.js barely registers.
Verbal Identity
Verbal identity is the complete ruled language system that dictates a brand's voice, vocabulary, tone, and writing mechanics before any visual work begins.
Vercel Preview
Vercel Preview is the automatic live URL that deploys from every pull request and turns your branch into a real review surface instead of another screenshot lie.
Vertical Rhythm
Vertical rhythm is the consistent vertical spacing of text, paragraphs, and UI elements aligned to a baseline grid that creates predictable flow and visual harmony across your interfaces.
Vibe Coding
Vibe coding is designers using natural language to describe the exact feeling behavior and references they want then watching AI generate real React Tailwind or shadcn code instead of static Figma frames.
Vibe First
Vibe First tools prioritize the magical feeling of instant app creation and streaming previews over clean exportable code and production ownership. Lovable defined the stance in 2026 by hiding every developer detail behind plain English chat while the live app updated in real time.
Vibes Export
Code output from an AI vision model that visually approximates a design but lacks fidelity to its underlying system, semantic structure, or maintainability.
Video First Hero
Hero pattern that uses an autoplaying muted video as the primary demo element. The video is not decoration. It is the product when motion and category are identical.
View Transitions
The web API that captures two page states and animates the swap so navigation feels like a DOM update instead of a full reload.
Vision Stack
The vision stack is the set of AI models designers route screenshots to for analysis, token extraction, and code generation. It matches each image to the model best suited for the task whether that is Claude for design reasoning Cursor for implementation or Gemini for competitive synthesis.
Visitor Temperature
Visitor temperature measures how much belief a new arrival already carries about your product. Cold visitors need proof, warm visitors need clarity, and hot visitors need the path cleared of any friction.
Visual CMS
A visual CMS gives content teams a browser-based editor that mirrors the live site so they can update pages, swap images, and publish collections without writing code or waiting on developers.
Visual Debt
Visual debt is the growing pile of design inconsistencies, awkward lockups, mismatched systems, and permanent workarounds that accumulate when brand architecture fails to match the reality of your expanding portfolio.
Visual Diff
Automated pixel regression that flags any unintended visual change against a locked baseline using Playwright screenshots, Pixelmatch comparison, and Chromatic review.
Visual Hierarchy
The arrangement of design elements so the eye processes them in a deliberate order, controlled by size, contrast, color, spacing, and position.
Visual Identity
The visible elements of a brand: logo, color palette, typography, imagery style, and layout patterns.
Visual Moat
The unique visual language a product owns across every surface that creates instant recognition and makes replication expensive and obvious.
Visual QA
Visual QA is the structured audit that compares production code against the four-layer Figma file to catch every token mismatch, variant error, and breakpoint failure before users see it.
Visual System
A visual system is the complete set of rules assets and behaviors that translate brand strategy into consistent executable design decisions across every touchpoint.
Visual Taste
Visual taste is the designer's gut judgment of whether a layout or detail simply feels right. Models like Claude Opus 4.8 can swallow your entire design system yet still cannot make this call.
Voice Attributes
Voice attributes are the three to five precise personality traits that define your brand's consistent character so every writer, designer, and AI prompt produces copy that actually sounds like you.
Voice Audit
A voice audit is the brutal inventory of every piece of copy your brand has shipped so you can spot exactly where you sound like yourself and where you sound like every other corporate zombie.
Voice JSON
Voice JSON is a structured data block that encodes your brand voice attributes, never-say list, tone contexts, and side-by-side rewrites into one pasteable object any AI can consume without guessing.
Voice Rubric
A voice rubric turns brand voice from vague adjectives into measurable rules and scores that AI and humans can follow consistently at high volume.
W
Walk-away Price
Walk-away price is the hard floor below which you decline the project. Calculate it as your quote floor minus 10 percent so you never subsidize a client's business with your own burnout and lost opportunity.
Warm Traffic
Warm traffic consists of visitors who already know and trust your brand before they land on your site, arriving via direct visits, branded searches, email links, or trusted referrals.
WCAG
The Web Content Accessibility Guidelines, published by W3C, defining measurable criteria for making digital content usable for people with disabilities, including color contrast, keyboard navigation, and screen reader support.
Weight and Style Rules
Weight and style rules are the non-negotiable directives dictating how every piece of text in your design system uses different font weights and styles. They establish visual hierarchy, guide the reader's eye, and communicate meaning consistently across all brand touchpoints.
Weight Axis
The weight axis packs every thickness from 100 to 900 inside one variable font file so you control font-weight like 350 or 723 through CSS while the browser interpolates the outlines in real time.
Weight Hierarchy
Weight hierarchy assigns a fixed font weight to every text role in your system so bold always signals importance, medium always supports, and regular always carries content.
White Space
The empty area between and around design elements that creates breathing room, establishes hierarchy, and improves readability.
Whitespace
Whitespace is the empty area in a composition that shapes how the eye moves, groups elements, and signals hierarchy without adding any visual content.
Whitespace Discipline
Whitespace discipline is the senior habit of designing the empty space first then dropping every element into gaps that were carved on purpose. It keeps the exact same air-to-ink ratios from marketing hero to dashboard card so the brand never contradicts itself.
Width Axis
The variable font axis that condenses or expands letterforms via font-stretch percentages, perfect for responsive editorial layouts from a single file.
Wireframe
A deliberately low-fidelity layout sketch that locks structure, hierarchy, and content placement before any visual design or interaction polish is applied.
Wordmark
A wordmark sets the full brand name in distinctive typography to function as the complete logo without symbols or abbreviations.
Workhorse Font
A workhorse font is a highly legible typeface engineered for body copy and sustained reading at 14 to 18 pixels. It forms the quiet foundation of any system that ships while display fonts handle the headlines.
Working Artifact
A functional, interactive design deliverable, often AI-generated, that replaces static specifications by communicating intent through live behavior and inspectable code.
Working Memory
Working memory is the AI's active context window, the exact slice of tokens from your chat history, files, tool outputs, and instructions that the model can actually see and process for its next response.
Workstream
A workstream is one focused AI chat dedicated to a single objective so the model receives clean relevant context instead of a noisy mix of unrelated tasks.
Written Spec
The one-page document that names the problem, exact user segment, single success metric, failure modes, and cut lines before anyone opens Figma. It is the primary artifact of the 2026 designer-PM role with the Figma file relegated to appendix status.
Z
Z-Pattern
An eye-tracking pattern where users scan pages in a Z-shape: across the top, diagonally to the bottom-left, then across the bottom. Common on pages with minimal text and clear CTAs.
Zero Data Retention
Zero Data Retention means the AI provider deletes your prompts, uploaded design files, outputs, and full conversation the instant the response finishes. Nothing is logged, backed up, or held for 30 days.
Zero Handoff
Zero handoff removes the designer-engineer wall entirely so designers ship production code and engineers critique visual decisions in the same loop. The practice collapses traditional handoff friction into shared ownership that produces higher fidelity at higher speed.
Zero State
The empty state a new user sees on their very first visit before any data has been created. It focuses specifically on initial no-data situations rather than filtered results or permission errors.
Zero-to-one design
The initial design phase focused on rapidly generating a tangible first draft from a blank slate, often leveraging AI to bypass traditional design tool complexities.