ScrapeGraph MCP Server

Name: ScrapeGraph MCP Server
Author: ScrapeGraphAI

@ScrapeGraphAI

代码仓库

关于 ScrapeGraph MCP Server

ScapeGraph MCP Server

基本信息

分类

其他

许可证

MIT

运行时

python

传输方式

stdio

发布者

ScrapeGraphAI

配置

使用下面的配置,将此服务器添加到你的 MCP 客户端。

{
  "mcpServers": {
    "scrapegraph-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@smithery/cli",
        "install",
        "@ScrapeGraphAI/scrapegraph-mcp",
        "--client",
        "claude"
      ]
    }
  }
}

工具

POST /scrape (`output_format`: markdown, html, screenshot, branding, links, images, summary)

POST /extract (requires `website_url` + `user_prompt`; optional `output_schema`)

POST /search (`num_results` 1–20; supports `country_search`, `time_range`, `output_schema`)

POST /crawl — `extraction_mode` markdown / html / links / images / summary / branding / screenshot

GET /crawl/:id (poll until `status: completed`)

POST /schema (generate or augment a JSON Schema from a prompt)

GET /credits

GET /history (paginated, `service` filter)

GET /monitor/:id/activity (paginated tick history: `id`, `createdAt`, `status`, `changed`, `elapsedMs`, `diffs`)

概览

What is ScrapeGraph MCP Server?

The ScrapeGraph MCP Server is a production-ready Model Context Protocol (MCP) server that integrates with the ScrapeGraph AI API. It enables language models to perform AI-powered web scraping tasks such as scraping, extracting structured data, searching, crawling multi-page sites, generating schemas, and managing monitors—all with enterprise-grade reliability.

How to use ScrapeGraph MCP Server?

Obtain a ScrapeGraph API key from the ScrapeGraph Dashboard. Then install via Smithery (npx -y @smithery/cli install @ScrapeGraphAI/scrapegraph-mcp --client claude) or manually for local use. Configure the server in Claude Desktop, Cursor, or Google ADK with your API key. The server communicates over stdio or HTTP (remote) and exposes tools like scrape, extract, search, crawl_start, monitor_create, and credits.

Key features of ScrapeGraph MCP Server

Scrape webpages in multiple output formats (markdown, HTML, screenshot, etc.)
Extract structured data using AI prompts and optional JSON schemas
Perform web searches with configurable country and time filters
Asynchronous multi-page crawling with start, status, stop, and resume
Generate and augment JSON schemas from prompts
Manage scheduled monitoring jobs with activity history
Check account credits and paginated history

Use cases of ScrapeGraph MCP Server

Convert a webpage to markdown for summarization
Extract all product prices and descriptions from an e-commerce site
Research the latest AI developments and summarize findings
Monitor a set of pages for changes over time
Crawl an entire documentation site to extract structured data

FAQ from ScrapeGraph MCP Server

What runtime does the ScrapeGraph MCP Server require?

Python 3.13 or higher is required for local usage. For remote server usage, only Node.js is needed to run the MCP remote client.

How do I get an API key for ScrapeGraph AI?

Sign up and generate your API key from the ScrapeGraph Dashboard. Set it as the SGAI_API_KEY environment variable or pass it via the scrapegraphApiKey or X-API-Key header.

Can I use the server without installing anything locally?

Yes, you can connect to the hosted remote MCP server at https://mcp.scrapegraphai.com/mcp by configuring your MCP client with the URL and your API key.

What tools are available in the latest v3 version?

Tools include scrape, extract, search, crawl_start, crawl_get_status, crawl_stop, crawl_resume, schema, credits, history, monitor_create, monitor_list, monitor_get, monitor_pause, monitor_resume, monitor_delete, and monitor_activity.

What should I do if tools are not appearing in my AI assistant?

Check Claude Desktop logs (macOS: ~/Library/Logs/Claude/, Windows: %APPDATA%\Claude\Logs\), verify the server starts without errors, and ensure your configuration JSON is valid. Also confirm that Python and the package are properly installed.