If you spend your day inside Claude, Cursor, or Windsurf, switching over to a terminal or a Playground tab just to run a scrape breaks your flow. You end up copying URLs back and forth, pasting JSON into your editor, and losing the thread of whatever you were actually working on.
MCP is the standard that lets an AI assistant reach outside the chat and use real tools.
The Spidra MCP server puts scraping, batch processing, and crawling behind that standard, so the assistant can pull live web data into the same conversation instead of you fetching it yourself.
The server exposes 12 tools covering everything the API does: scraping single pages, batch processing up to 50 URLs in parallel, crawling entire sites, and checking on jobs that outlive their wait window.
Ask your assistant to scrape a page, and it calls spidra_scrape directly:
{
"name": "spidra_scrape",
"arguments": {
"urls": ["https://example.com/product"],
"prompt": "Extract the product name, price, and description",
"output": "json",
"schema": {
"type": "object",
"properties": {
"name": { "type": "string" },
"price": { "type": "number" },
"description": { "type": "string" }
},
"required": ["name", "price"]
}
}
}
The server is open source at github.com/spidra-io/spidra-mcp-server, MIT licensed, and published as spidra-mcp on npm.
Read the full docs here.
