Cue

Cue is an Awwwards-tier UI component library for anyone building to stand out. It collects best-in-class components from across the internet, curated by hand, so that designers, developers, agencies, solo makers, and product teams who refuse to ship generic-looking work can bring high-end web design closer to their next project. Every entry is a component reference sourced from an Awwwards Site of the Day, a Behance-featured interaction, or a production output the founder considered best-of-class. Cue does not host projects, deploy code, or run an AI model itself; it is the taste layer on top of the tools people already use. The problem Cue addresses is a familiar one for builders: a website can function perfectly and still look entirely forgettable. Bold hero sections, smooth interactions, unique layouts, and thoughtful micro-animations are the elements that make a site stand out, but finding genuinely best-in-class examples and then reproducing them takes time, taste, and a lot of searching. The strongest references are scattered across Awwwards, CollectUI, and X, presented as finished sites rather than as reusable, copyable pieces. Cue gathers those references in one place and pairs each one with the code and prompt needed to recreate it in your own project, shortening the distance between spotting a great interaction and shipping something that feels like it. The core of Cue is its growing collection of hand-picked components — 75+ and counting, with new drops arriving daily. Each entry is a component reference rather than a machine-generated filler item, sourced from an Awwwards-tier website or a best-in-class interaction. Cue explicitly states that nothing is machine-generated to fill the grid; every drop is hand-picked by the founder, and there is a dedicated editorial section highlighting the new drop of the day. The library covers bold hero sections, smooth interactions, unique layouts, and thoughtful micro-animations, organised so builders can browse sections and interactions separately and sort between new and old entries. Because the bar is taste rather than volume, if a user finds a better reference for the same component, the founder replaces it. Every Cue component ships with an AI prompt designed to be dropped directly into the AI builder or assistant you already use. The prompts are written for tools including Bolt, v0, Cursor, Framer AI, ChatGPT, and Claude, so a reference you like can be turned into working output inside your existing workflow rather than being rebuilt by hand. For builders working in AI-native environments, this reframes the component library as an input to their tools rather than a separate destination. Free members can copy two AI prompts per 24 hours, which allows anyone to test the format before committing to a paid plan. Alongside prompts, Cue is actively rolling out React source code and an MCP server for its components. The MCP server is described as native access from Cursor, Claude Desktop, and any MCP-aware AI tool, meaning components can be pulled into an AI-assisted development session without leaving the editor or assistant. React source code gives developers a concrete implementation to adapt rather than a description to interpret. Together these two channels cover builders who prefer to prompt, builders who prefer to code, and builders who work in a mix of both — while the website itself stays a reference layer rather than a hosting or deployment platform. Browsing is built for fast discovery. A ⌘K command search lets you jump straight to what you need, and the library can be filtered by tags covering sections and interactions, sorted between new and old entries, and filtered by price. A 'new drop today' area highlights what has just been added, and visitors can subscribe by email to be told when new components ship — described as no spam, unsubscribe anytime. There is also a Discord community where members help pick the next drop. The overall approach is deliberately lightweight: Cue is explicitly positioned as not a template pack, not a subscription, not a course, not an AI wrapper, and not a marketplace. For users, the outcome is a shorter path from inspiration to implementation. Instead of bookmarking a striking site and then reverse-engineering it, you get the reference, the prompt, and — as it rolls out — the React code and MCP access. The stated promise is simple: build less, create more. That means less time hunting for reference material, less time translating a visual idea into an implementation, and a result that does not look generic. For teams and freelancers whose work is judged on visual quality, Cue acts as an external taste layer that raises the baseline of what they ship. Typical scenarios revolve around building websites that stand out. A solo maker preparing a landing page can pull a hero section reference and drop its prompt into v0 or Bolt. A developer working in Cursor or Claude Desktop can reach the same components through the MCP server as it rolls out. A designer looking for a specific interaction can browse the interactions tag, find a reference sourced from an Awwwards Site of the Day, and use the attached prompt to recreate it. Agencies and product teams can use the curated set to raise the visual bar across client projects, and founders can subscribe to email drops to see each new component as it ships or join Discord to help choose the next one. Cue is explicitly not only for AI builders. It is aimed at designers, developers, agencies, solo makers, and product teams who refuse to ship generic-looking work, and the Product Hunt listing places it in Design Tools, Developer Tools, and Web Design. Pricing spans several options. A Free tier lets you browse the library with two AI-prompt copies per 24 hours. Cue+ Founding Lifetime is USD $99 one-time, capped at the first 50 members. Cue+ Lifetime standard is USD $249 one-time after the founding tier sells out. A time-limited offer shows $99 marked down to $79 lifetime with coupon code CUE49, and a custom pricing option lets builders pay only for the components they pick. The founder also offers a paid service to build a site that stands out. On the integration side, Cue names Cursor, v0, Bolt, Framer AI, ChatGPT, Claude, Claude Desktop, and any MCP-aware AI tool. In short, Cue is a hand-curated, Awwwards-tier component library paired with AI prompts, React source, and MCP access — a taste layer that helps builders move from a best-in-class reference to a shipped interface without settling for generic-looking work.