Google Search and Google Shopping MCP server

Step by step

1. 01Know the connection details

   Server URL: https://api.crawlerdata.com/mcp. Transport: Streamable HTTP (stateless, no session to keep). Authentication: sign in with your CrawlerData account (OAuth: the client opens a browser window), or send an "Authorization: Bearer YOUR_API_KEY" header on every request.

   Tools: google_shopping_search, google_shopping_sellers, google_shopping_product, google_search, google_images, google_news, google_local and google_autocomplete. The server is also listed in the official MCP Registry as com.crawlerdata/google-shopping-serp.

2. 02Claude Code

   Run this in your project. --transport http selects Streamable HTTP and --header sets the auth header. Add --scope user to make the server available in every project.

   claude mcp list should then show crawlerdata as Connected.

   Prefer to sign in instead of pasting a key? Add the server without --header (claude mcp add --transport http crawlerdata https://api.crawlerdata.com/mcp), run /mcp in Claude Code, choose crawlerdata and Authenticate, then log in to CrawlerData in the browser and click Allow.

   claude mcp add --transport http crawlerdata https://api.crawlerdata.com/mcp \
     --header "Authorization: Bearer YOUR_API_KEY"
   
   # Check it connected
   claude mcp list

3. 03Claude Code, shared with a team (.mcp.json)

   To check the server into a repository without the key, put this in .mcp.json at the project root. Claude Code expands ${CRAWLERDATA_API_KEY} from each person's environment, and asks each person to approve project servers the first time.

   {
     "mcpServers": {
       "crawlerdata": {
         "type": "http",
         "url": "https://api.crawlerdata.com/mcp",
         "headers": {
           "Authorization": "Bearer ${CRAWLERDATA_API_KEY}"
         }
       }
     }
   }

4. 04Claude Desktop and claude.ai: custom connector

   Open Connectors (under Customize, or Settings in older versions), choose Add, then Add custom connector, enter the server URL and click Add. Claude opens a CrawlerData sign-in window: log in and click Allow, and the tools are ready. Calls are billed to your organization's credits. To use an API key instead, choose No sign in and add a request header Authorization with the value Bearer YOUR_API_KEY.

   Custom connectors are available on Free, Pro, Max, Team and Enterprise plans (Free allows one). On Team and Enterprise plans an owner adds them in the organization settings.

5. 05Claude Desktop: claude_desktop_config.json

   If you prefer a local config file, Claude Desktop starts local (stdio) servers from claude_desktop_config.json, so use the mcp-remote bridge to reach the remote server. Open the file from Settings, then Developer, then Edit Config, add the block below and restart Claude Desktop.

   There is no space after "Authorization:" on purpose. mcp-remote's documentation notes that Claude Desktop on Windows and Cursor do not escape spaces in args, so the space-containing value goes in env instead. This form works on macOS too.

   {
     "mcpServers": {
       "crawlerdata": {
         "command": "npx",
         "args": [
           "-y",
           "mcp-remote@0.14.3",
           "https://api.crawlerdata.com/mcp",
           "--header",
           "Authorization:${AUTH_HEADER}"
         ],
         "env": {
           "AUTH_HEADER": "Bearer YOUR_API_KEY"
         }
       }
     }
   }

6. 06Cursor

   Add the server to .cursor/mcp.json in your project, or ~/.cursor/mcp.json for every project. Cursor reads ${env:CRAWLERDATA_API_KEY} from your environment, so the key stays out of the file. You can also paste the key in place of the variable.

   {
     "mcpServers": {
       "crawlerdata": {
         "url": "https://api.crawlerdata.com/mcp",
         "headers": {
           "Authorization": "Bearer ${env:CRAWLERDATA_API_KEY}"
         }
       }
     }
   }

7. 07Check the server from a script

   Any client that speaks JSON-RPC over HTTP can use the server. This script initializes, lists the tools and calls google_autocomplete (1 credit). It is a quick way to confirm your key works before you debug a client config.

   """Talk to the CrawlerData MCP server directly: initialize, list tools, call one. Needs requests."""
   import json
   import os
   
   import requests
   
   URL = "https://api.crawlerdata.com/mcp"
   HEADERS = {
       "Authorization": f"Bearer {os.environ['CRAWLERDATA_API_KEY']}",
       "Accept": "application/json, text/event-stream",
   }
   
   
   def rpc(id, method, params):
       body = {"jsonrpc": "2.0", "id": id, "method": method, "params": params}
       resp = requests.post(URL, headers=HEADERS, json=body, timeout=200)
       resp.raise_for_status()
       return resp.json()["result"]
   
   
   info = rpc(1, "initialize", {"protocolVersion": "2025-06-18", "capabilities": {},
                                "clientInfo": {"name": "check", "version": "1.0"}})
   print(info["serverInfo"])
   
   tools = rpc(2, "tools/list", {})["tools"]
   print([t["name"] for t in tools])
   
   call = rpc(3, "tools/call", {"name": "google_autocomplete", "arguments": {"query": "running sho"}})
   data = json.loads(call["content"][0]["text"])
   print(data["credits"], [s["value"] for s in data["result"]["suggestions"][:3]])

8. 08Ask for something

   Try prompts such as "Find the Google Shopping product for the Sony WH-1000XM5 and list every seller with total price", "What does Google show for best crm software in the UK?" or "Get the latest Google News results for nvidia". The assistant picks the tool and the parameters. Tool results include the credits each call used.

If something goes wrong

The client shows the server as failed or 401

Check the header is exactly "Authorization: Bearer YOUR_API_KEY". In the mcp-remote config, keep "Authorization:${AUTH_HEADER}" with no space and put "Bearer YOUR_API_KEY" in env.

Claude Desktop does not show the tools

Restart Claude Desktop fully after editing claude_desktop_config.json, and check that node and npx run from a terminal (Node 18 or newer). Claude Desktop uses the Node.js on your system PATH.

402 INSUFFICIENT_CREDITS

Your balance is too low for the request. Buy a credit pack in the dashboard. Requests that fail for other reasons are not charged, so a 402 never costs credits.

429 RATE_LIMITED

Each API key allows 100 requests per minute and 20 per second. Wait until the time in the X-RateLimit-Reset header (a Unix timestamp) and retry, or spread requests out.

A sellers call times out in the client

google_shopping_sellers collects every seller on the product page, so it needs more time than a search. Raise the client's tool timeout if it has one, or ask for google_shopping_product (details only) when you do not need every seller.

Request and response schemas: OpenAPI 3.1 document. Still stuck? Email contact@crawlerdata.com.

APIs you can call

Solutions built on it

Other integrations