Vogue MCP Server

A comprehensive Model Context Protocol (MCP) server for extracting fashion intelligence from Vogue. Built for AI Agents like Claude Desktop, Cursor, and Windsurf to seamlessly pull runway images, trend articles, and styling reference shots.

Features

• 🕵️ Stealth Browser Engine: Uses Playwright + puppeteer-extra-plugin-stealth in headless mode to bypass anti-bot and "Download App" interceptwalls.
• 🎨 Base64 Vision Injection (inspect_image_visually): Eliminates AI hallucination by shrinking 4K Vogue CDN images to 768px and returning them as MCP Image Blocks directly to the Agent's visual cortex.
• 🔎 Full Search (search_vogue): Query Vogue's internal engine for articles and galleries.
• 📸 High-Res Runway Extraction (get_collection_looks): Pull unwatermarked, raw runway photos by season and brand.
• 📖 Editorial + Gallery Extraction (get_article_content): Extracts structured text from articles and full image blocks from Vogue slideshow/gallery pages such as street style coverage.
• 💃 Street Style (get_street_style_galleries): Automatically hooks into bundle containers to extract the latest street style coverage.

Quick Start (NPM)

If your Agent supports npx commands (like Cursor or Claude Desktop), simply add this configuration without needing to clone anything:

{
  "mcpServers": {
    "vogue-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@pigbun-ai/pigbun-vogue-mcp"
      ],
      "env": {
        "PIGBUN_API_KEY": "your_api_key_here"
      }
    }
  }
}

Authentication

This MCP server requires a valid PIGBUN_API_KEY to function. You can obtain one by registering at the PigBun AI Dashboard. Without this API key, the local proxy will prevent data extraction.

Available Tools

1. search_vogue
2. get_designer_directory
3. get_brand_collections
4. get_collection_looks
5. get_collection_metadata
6. get_street_style_galleries
7. get_article_content
8. inspect_image_visually
9. download_images_to_local
10. login

Private Repository Details

Maintained by PigBun-AI.

Development

Execute npx ts-node src/test-scraper.ts locally to run the end-to-end extraction pipeline manually.