Scrape Everything
Quickstart with MCP / CLIMenu

Quickstart with MCP / CLI

Two ways to use the API without writing an HTTP client: from your AI assistant through MCP, or from your terminal with the CLI. Both use the same API key and credits.

1. Get an API key

  1. Continue with Google. New accounts get 100 free credits.
  2. Create a key on the API keys page and store it as SCRAPEEVERYTHING_API_KEY.

2a. Connect your AI assistant (MCP)

Claude Code:

Terminal
claude mcp add scrapeeverything -e SCRAPEEVERYTHING_API_KEY=ssk_live_... -- npx -y @scrapeeverything/mcp

Claude Desktop, Cursor or Windsurf: add this to the MCP config file.

mcp config
{
  "mcpServers": {
    "scrapeeverything": {
      "command": "npx",
      "args": ["-y", "@scrapeeverything/mcp"],
      "env": {
        "SCRAPEEVERYTHING_API_KEY": "ssk_live_..."
      }
    }
  }
}

Then ask, for example: “Get the TikTok profile of @khaby.lame”. Full setup for every client, including VS Code, is on the MCP page.

Slow calls: our MCP server waits up to 120 seconds per call, and a request can run for about 75 seconds before we stop it and refund it. If your client gives up sooner, raise its tool timeout (in Claude Code, a per-server timeout in .mcp.json or MCP_TOOL_TIMEOUT). Calls that fail or time out on our side are refunded. More on timeouts.

2b. Use the terminal (CLI)

Terminal
export SCRAPEEVERYTHING_API_KEY=ssk_live_...
npx scrapeeverything instagram profile --username nasa
npx scrapeeverything youtube transcript --url "https://www.youtube.com/watch?v=jNQXAC9IVRw"

More commands and output options are on the CLI page.

Next steps

Add ready-made workflows with Agent Skills, or import the OpenAPI spec into your tools. Credit costs are on each endpoint reference page.