The fastest web crawling, scraping, search, and browser automation for AI agents — delivered through the hosted Spider Cloud MCP server, with built-in anti-bot bypass, proxy rotation, and AI-driven extraction.
- Website: https://spider.cloud
- MCP docs: https://spider.cloud/mcp
- Get an API key: https://spider.cloud/api-keys
- 1 MCP server — the hosted Spider endpoint at
https://mcp.spider.cloud/mcp, exposing crawl, scrape, search, links, screenshot, unblocker, transform, credits, AI extraction, and remote browser-automation tools. - 3 skills —
spider(overview + tool-selection guide),spider-crawl-scrape(crawling, scraping, search, bot bypass), andspider-browser(interactive browser sessions).
The Spider MCP server uses OAuth — on first connection the agent prompts you
to authorize in the browser, with no API key to paste. The bundled
.mcp.json just points at the hosted endpoint:
{
"mcpServers": {
"spider": {
"type": "http",
"url": "https://mcp.spider.cloud/mcp"
}
}
}The endpoint advertises OAuth via protected-resource metadata with dynamic client registration and PKCE, so MCP clients authorize automatically.
API-key alternative. The server also accepts a Bearer API key if you prefer
not to use OAuth — add an Authorization header to the config:
"headers": { "Authorization": "Bearer ${SPIDER_API_KEY}" }with SPIDER_API_KEY from https://spider.cloud/api-keys.
Core tools run on pay-per-use credits (no subscription required). The AI tools require an AI subscription.
- Endpoint called:
https://mcp.spider.cloud/mcp(the official Spider Cloud hosted MCP server) and its OAuth discovery/authorization endpoints under the same host. No other network endpoints. - Credential: none stored by the plugin. Auth is OAuth (browser authorization, tokens managed by the MCP client) or, optionally, a user-supplied
SPIDER_API_KEYsent only as theAuthorization: Bearerheader to the endpoint above. Nothing is read from the filesystem and no telemetry is collected by this plugin.
MIT — see LICENSE.