MCP Skill

The scraperapi‑mcp skill teaches your LLM agent how to make the most of the ScraperAPI MCP tools. Installing the skill provides best-practice guidance, automatic tool selection, and recipes for common web‑data workflows. Without it, you can still call the MCP tools by using the scrape keyword and specifying parameters like url, country_code, render, etc.

The skill helps your agent choose between the various ScraperAPI tools (scrape, google_search, amazon_search, etc.) based on the task. For example, it suggests using google_search for research or amazon_product for product details.

Installation

Prerequisites:

  • ScraperAPI account.

  • Node.js installed to run npx commands.

  • A client that can load agent skills (for example Claude Desktop with MCP enabled).

# Install directly from GitHub
npx skills add scraperapi/scraperapi-skills

How it Works

When your agent detects that a task matches the skill’s triggers (scraping, Google/Amazon/Walmart/eBay searches, Redfin queries, crawling, etc.), it loads the skill’s guidance into the system prompt. The skill contains:

  • A list of all available ScraperAPI tools and their parameters.

  • A decision tree for selecting the right tool.

  • Best practices for cost, pagination, localization and result formatting.

  • Guidance on handling large outputs efficiently.

Prompt Example

Real-time News & Market Data

- "I need latest Google news for the current price of Gold."

Response

Claude now factors in the SKILL.md file when making a decision. Instead of manually parsing the HTML or relying on raw search results, it sends a request to ScraperAPI's Google News Structured Data endpoint. That endpoint is specifically designed to return clean, structured JSON data for Google News results

ScraperAPI handles:

  • Bypassing the Bot protection mechanisms.

  • Fetching the Google News results.

  • Structuring the response into machine-readable data.

chevron-rightGoldPriceResponse.mdhashtag

Last updated