search_designs

Search the catalog by category, tag, or keyword.

Fuzzy-search the catalog by name, tagline, tags, or categories. Returns up to 20 matches sorted by relevance — use this when the user says “something like Linear” or “find me dark editorial designs.”

Input schema

{
  query: string;       // Free-text search term (required)
  limit?: number;      // Max results, default 20, max 50
}

query is required. The search is fuzzy — it matches character sequences, not whole words. “lnr” will match “Linear,” “dark” will match designs tagged with dark, “fintech” will match designs in the fintech category.

Output

{
  query: string;
  count: number;
  matches: [
    {
      slug: string;
      name: string;
      tagline: string;
      tags: string[];
    }
  ]
}

Each match includes the design’s slug (for use with get_design), display name, tagline, and tags.

Example: agent dialogue

Scenario 1: Keyword search

User: “Find me something like Linear but darker.”

Agent calls search_designs({ query: "linear dark" })

Returns: Linear, Vercel, Notion, Figma (matches on name “linear”, tag “dark”)

Agent: “Linear, Vercel, and Figma all have dark-canvas designs. Vercel has the most minimal approach. Let me pull Vercel’s exact tokens…”

Scenario 2: Category search

User: “Show me fintech designs with minimal color palettes.”

Agent calls search_designs({ query: "fintech minimal" })

Returns: Robinhood, Stripe, Wise (matches on category fintech, tag minimal)

Agent: “Stripe and Robinhood are fintech with minimal color palettes. Stripe uses a restrained blue-and-gray system…”

Scenario 3: Tag search

User: “What B2B SaaS brands use editorial typography?”

Agent calls search_designs({ query: "editorial b2b saas" })

Returns: Intercom, Slack, Notion (matches on tags editorial, categories b2b + saas)

Agent: “Intercom and Slack both emphasize editorial type scales. Intercom uses a serif system for contrast…”

Search semantics

The fuzzy matcher searches across:

  • Name — “linear” matches “Linear”, “stripe” matches “Stripe”
  • Tagline — matches short taglines on file
  • Tags — “dark”, “minimal”, “editorial”, “glassmorphism”, etc.
  • Categories — “fintech”, “dev-tools”, “ai”, “saas”, “media”, etc.

The search is case-insensitive and character-sequence-based (not word-based). “lnr” matches “Linear,” “str” matches “Stripe.” Matches are returned in catalog order (not ranked by relevance).