performative-ui

repository·main·Indexed 21 days ago

https://github.com/vorpus/performativeui

A library of AI-native React components for React 19 and TypeScript, designed to signal the level of oversubscription in a funding round. The project includes research on performative AI UI design systems, covering patterns for typewriter hero animations, logo walls, and decorative backgrounds used by AI companies.

Tokens
19.1K
Snippets
33
Records
98
Agent score
73%

What's inside performative-ui

  1. Reference canonical AI landing pages for performative UI research

    main

    The 00_source_companies.md file serves as a master catalog of AI companies used for performative AI UI design-system research. It categorizes companies into functional groups (e.g., Frontier labs, Coding tools, Vertical SaaS, etc.) to provide canonical examples of the "performative AI UI" aesthetic.

    Developers or design agents can use this roster to identify industry-standard UI patterns across different sectors like:

    • Frontier Labs: OpenAI, Anthropic, Mistral.
    • Coding/Dev Tools: Cursor, Replit, v0.
    • Productivity/Search: Perplexity, Notion AI.
    • Vertical SaaS: Harvey (Legal), Hippocratic AI (Healthcare).
    • Infrastructure/RAG: Pinecone, LangChain.
    • Generative Media: Runway (Video), Midjourney (Image), ElevenLabs (Voice).
  2. Identify AI company categories for UI analysis

    main

    When performing UI research or analysis using this repository's source list, companies are classified using the following standard categories:

    • LLM app
    • Agents
    • AI infra
    • Dev tools
    • Coding
    • Voice
    • Video
    • Image
    • Vector DB
    • RAG
    • Eval/Observability
    • Vertical SaaS (often qualified by domain, e.g., Legal, Healthcare, Sales, Finance, HR, Education, Compliance, Real Estate, Construction, Supply chain, Marketing, EdTech, Chem)
    • Robotics
    • Vision
    • Consumer AI
    • Data/Training
    • Audio / Generative
    • 3D / Generative
    • Biotech AI
  3. Implement a Single-Letter Morph effect

    main

    The Single-Letter Morph is a hero effect where a static sentence contains one specific character that cycles through different Unicode glyphs, character-set variants (like fraktur or script), or glitch variants. This is often used by generative AI brands to signal a 'liquid' or generative brand identity.

    Example Shape: tr𝒶nsform where the a cycles through variants.

  4. Implement Logo Wall Visual Treatments

    main

    When designing a logo wall, choose a visual treatment based on the desired brand signal:

    • Grayscale monochrome row: The standard, professional default.
    • Infinite horizontal scroll marquee: Signals high volume/many customers. Typically implemented using CSS @keyframes scroll with translate3d and duplicating the logo strip 2-3x for a seamless loop. Common libraries include react-fast-marquee or Framer Motion's <motion.div>.
    • Dark-bg white-logo carousel: A "premium" aesthetic (e.g., Sierra, 11x).
    • Color logos on light bg: Signals high confidence (e.g., Glean).
    • Logo + per-company metric tiles: A modern trend (post-2024) that combines the logo with specific usage data (e.g., Modal, Braintrust).
    • Logo + customer-quote-with-headshot: The "humanized" version (e.g., Cursor, Replit).
  5. Use AI Iconography: The Sparkle and Star Clusters

    main

    To signal AI capabilities visually, use specific iconography patterns:

    • The Sparkle ( / Four-Pointed Star): Place a 16px four-pointed star to the left of any verb meaning 'generate' (e.g., 'Generate', 'Ask AI'). Use white-on-gradient or gradient-on-white color schemes. Avoid red, as it implies an error.
    • The Stars-in-a-Cluster (Magic Wand variant): Use a cluster of three sparkles in descending size to signal a broad category or 'intelligence' section, rather than a single feature.
  6. Implement AI Backgrounds: Aurora and Grids

    main

    Use these background patterns to establish an AI aesthetic:

    • The Aurora Gradient: Use slowly drifting bands of purple, pink, cyan, and orange behind the hero section to evoke a nebula or 'intelligence'.
    • The Black-with-Subtle-Grid: For a technical/developer-focused look, use a pure black background with a 1px gridded SVG at 4% opacity. This is often paired with code editor mockups.
    • Brand Imitations:
      • Anthropic-Orange: Use warm rust/amber tones (#D4A574-ish) to signal safety and alignment.
      • OpenAI-Green: Use #10A37F on interactive elements like send buttons to imply lineage to ChatGPT.
  7. Adopt the 'AI Aesthetic' color palette

    main

    To match the current 'AI startup' or 'v0' aesthetic, use the following color palette. This palette is characterized by dark backgrounds and high-saturation indigo/violet accents:

    • Background: #0a0a0a to #1a1a2e (near-black with a blue shift).
    • Primary accent: Indigo #6366f1 $\rightarrow$ Violet #8b5cf6 (Tailwind-style).
    • Secondary accent: Cyan #06b6d4 (for data signals) or Magenta #ec4899 (for creative AI signals).
    • Developer/Matrix vibes: Lime #84cc16.
    • Anthropic-style: Terracotta #d97757.
  8. Implement a Spotlight effect via CSS variables

    main

    A high-performance pattern for creating 'flashlight' or 'glow' effects (like the Stripe card grid) without triggering JS re-renders or animation loops.

    1. Attach a single pointermove listener to the container.
    2. On movement, update two CSS variables, --x and --y (using pixels or percentages), on the target element.
    3. Use these variables within a CSS radial-gradient for the background or a mask to reveal content.

    This approach is extremely lightweight (~10 lines of code) and is used by Stripe and Vercel.

    /* Example CSS usage */
    .card {
      background: radial-gradient(circle at var(--x) var(--y), rgba(255,255,255,0.2), transparent);
    }
  9. Implement smooth multiplayer cursor presence

    main

    When building multiplayer environments (the presence pattern), simply setting a cursor's position based on the latest server message results in jerky, 'teleporting' motion.

    To achieve smooth, continuous motion (as seen in Figma or tldraw), use interpolation. Instead of jumping to the new coordinates, use a short look-ahead spline or an interpolation library like perfect-cursors to smooth the path between server ticks.

  10. Implement a Scroll-Triggered Rotator

    main

    In this pattern, words do not auto-cycle based on a timer. Instead, the rotation is tied to the user's scroll position (often using tools like GSAP pinning). As the user scrolls, the page 'types in' a new noun or use case at specific scroll steps.

    Example Shape: A pinned section where each scroll increment triggers a new word in the typewriter sequence.

  11. Taxonomy of AI landing page background variants

    main

    This guide categorizes common visual patterns used in AI landing page design to signal 'intelligence' or 'complexity'. Use these patterns to decide on a visual direction for your project:

    • Particles / Node Graphs: Floating dots with lines connecting them (signals 'neural networks').
    • Constellation / Twinkling: Slower, shinier versions of particle graphs.
    • 3D Rotating Node Mesh: An icosahedron or similar 3D shape (signals 'knowledge graphs').
    • Wireframe Globe: A 3D spinning globe (signals 'global scale').
    • Topographic / Contour Lines: Heatmap-style lines (signals 'depth/data').
    • Gradient Orbs / Blurred Blobs: Large, soft, blurred divs (signals 'calm intelligence').
    • Stripe-style Gradient Mesh: WebGL shaders with noise/refraction (signals 'premium SaaS').
    • Animated Grid: A grid that warps near the cursor (signals 'spatial computing').
    • Floating Geometric Shapes: Abstract 3D forms (often exported from Spline).
    • Shader Noise: Organic, generative textures using fragment shaders.