Google Search Console MCP Server: No Google Cloud Setup

Most Google Search Console (GSC) MCP servers are open source, but they take a lot of effort to set up. You need to create a Google Cloud project, enable an API, build an OAuth client, download a credentials.json, then keep a local server running. PandaCrawl handles the Google connection for you. You sign up, connect your property, and start asking.

Setup: no Google Cloud project

1. Get PandaCrawl (a one-time licence).
2. Go through the onboarding. When prompted, click Connect Search Console and pick your property.
3. Add the PandaCrawl MCP endpoint to your client (steps below) and approve the PandaCrawl login when prompted.

The hosted MCP server URL is:

https://app.pandacrawl.app/mcp

Claude Code

claude mcp add --transport http --scope user pandacrawl https://app.pandacrawl.app/mcp

Claude Desktop

1. Open Settings → Connectors.
2. Click Add custom connector.
3. Paste https://app.pandacrawl.app/mcp.

Cursor

1. Open Cursor Settings → Tools & Integrations → MCP Tools.
2. Click New MCP Server (this opens mcp.json) and add:

{
  "mcpServers": {
    "pandacrawl": {
      "url": "https://app.pandacrawl.app/mcp"
    }
  }
}

Codex CLI

codex mcp add pandacrawl --url https://app.pandacrawl.app/mcp

Any other MCP client works the same way: add https://app.pandacrawl.app/mcp as a remote server and sign in. Full setup and troubleshooting live in the MCP docs.

The DIY way (every other GSC MCP)

1. Create a Google Cloud project
2. Enable the Search Console API
3. Configure the OAuth consent screen
4. Create OAuth client credentials
5. Download credentials.json
6. Install and run a local MCP server

Things to actually ask it

• "Which queries am I ranking on page two for that I'm close to pushing onto page one?"
• "Show me pages where impressions went up but clicks stayed flat — probably a title or meta problem."
• "Are any of my blog posts competing with each other for the same query?"
• "Is /pricing indexed? If not, what's blocking it?"
• "Which queries lost the most clicks over the last 28 days?"

What your agent can read

Two read-only tools over the property you connect. Both pull Google's own numbers, and neither uses credits.

Search performance. Clicks, impressions, CTR, and average position for your site, sliced by query, page, country, device, and date. Pull up to 16 months of history in a single question.

URL inspection. Whether a page is indexed and, if not, why: coverage status, last crawl, the canonical Google chose versus the one you declared, plus mobile and rich-result checks. Up to 10 URLs at a time.

The agent can read and inspect. It cannot change anything in your account.

What it costs

The Search Console tools cost nothing to run — Google doesn't charge you to read your own data, so reading it through PandaCrawl never draws on any balance. They come with the licence, alongside the tools that do query a paid provider (keyword research, rank tracking, backlinks), which run through the same MCP endpoint on your own DataForSEO account. PandaCrawl is also open source and can be self-hosted — see the FAQ below.

PandaCrawl vs. the alternatives

FAQ

What does it cost?

The Search Console MCP tools come with the PandaCrawl licence, a one-time purchase with a 30-day money-back guarantee, and cost nothing to run. PandaCrawl's other tools (keyword research, rank tracking, backlinks) run on your own DataForSEO account through the same endpoint, billed to you by DataForSEO at their rates. Self-hosting is free.

Is it open source?

Yes. PandaCrawl is open source, so you can self-host the whole thing for free, including the Search Console MCP. The one-click connection here uses PandaCrawl's hosted Google app. If you self-host, you bring your own Google OAuth client, the same Cloud-console step the hosted version saves you. Hosted means no setup; self-hosted means full control.

Is it read-only?

Yes. PandaCrawl requests read-only access (webmasters.readonly). Your agent can read performance data and inspect URLs, but it can't change your account.

Which AI clients work?

Any MCP client, including Claude Code, Claude Desktop, Cursor, Codex, and OpenClaw. Add the PandaCrawl MCP endpoint and sign in.

How fresh is the data?

It's Google's own data, so the most recent few days can be incomplete. History runs back 16 months, the same as the Search Console interface.