brightdata-mcp-adk-hackathon

概述

这是一个完整的网页抓取平台，结合了 BrightData 的 MCP 服务器和 Google 的代理开发工具包。它提供了包含 React 前端、FastAPI 后端和通过 BrightData 代理网络的企业级抓取功能的完整解决方案。该系统支持多种数据提取场景，包括电子商务、社交媒体和商业数据，由 Google Gemini 2.0 Flash 提供智能工具选择支持。

试试问 AI

装完之后，这里有 7 个你可以让 AI 做的事：

• search_engine

  Multi-engine search results from Google, Bing, Yandex

  试试:Search for the latest trends in artificial intelligence试试:Find information about sustainable energy solutions
• scrape_as_markdown

  Clean webpage extraction to markdown format

  试试:Extract the content from this webpage as markdown试试:Summarize the main points from this article
• scraping_browser_amazon

  Extract product information from Amazon

  试试:Find the top 10 selling laptops on Amazon试试:Get reviews and ratings for this product
• scraping_browser_walmart

  Extract product information from Walmart

  试试:Find the best deals on home appliances at Walmart试试:Check prices for this product across different categories
• scraping_browser_ebay

  Extract product information from eBay

  试试:Find vintage electronics on eBay试试:Get average selling price for this item
• scraping_linkedin

  Extract professional data from LinkedIn

  试试:Find tech companies in San Francisco on LinkedIn试试:Get information about job postings for software engineers
• scraping_instagram

  Extract social media data from Instagram

  试试:Find trending hashtags related to travel试试:Get information about this influencer's profile
• scraping_youtube

  Extract video data from YouTube

  试试:Find tutorials on machine learning试试:Get statistics for this popular video
• scraping_crunchbase

  Extract business data from Crunchbase

  试试:Find venture-backed startups in fintech试试:Get funding information for this company
• scraping_yahoo_finance

  Extract financial data from Yahoo Finance

  试试:Get stock market performance for tech companies试试:Find information about this company's earnings report
• scraping_google_maps

  Extract location data from Google Maps

  试试:Find restaurants near Times Square with high ratings试试:Get information about businesses in this area
• scraping_zillow

  Extract real estate data from Zillow

  试试:Find houses for sale in Seattle under $500,000试试:Get rental prices for apartments in downtown Chicago

可对比工具

安装

git clone https://github.com/arjunprabhulal/brightdata-mcp-adk-hackathon.git
cd brightdata-mcp-adk-hackathon

cp backend/config/.env.example backend/config/.env
# 编辑 .env 文件添加您的 API 凭据

docker compose up -d

{
  "mcpServers": {
    "brightdata-mcp": {
      "command": "python",
      "args": ["path/to/brightdata/mcp/server.py"],
      "env": {
        "BRIGHTDATA_API_TOKEN": "your_token_here"
      }
    }
  }
}

FAQ

这个平台可以抓取哪些类型的网站？

该平台支持抓取各种类型的网站，包括电子商务平台（亚马逊、沃尔玛）、社交媒体（领英、Instagram）、新闻来源、商业数据平台以及通过 BrightData 代理网络具有反机器人保护的自定义网站。

这个平台与其他网页抓取工具有什么不同？

该平台将 BrightData 的 MCP 服务器与 Google 的代理开发工具包和 Gemini 2.0 Flash AI 相结合，能够实现智能工具选择以实现最佳抓取效果。它提供包含 UI、API 和企业基础设施的完整解决方案，而不仅仅是抓取库。

速率限制和定价如何？

定价和速率限制取决于您的 BrightData 订阅计划和 API 使用情况。该平台包含用于管理 API 令牌和身份验证凭据的环境变量配置。

brightdata-mcp-adk-hackathon 对比