MCP Server · Model Context Protocol

The Google Merchant Center MCP Server

Give Claude, Cursor, Windsurf, Cline, Goose — or any MCP-compatible AI — direct access to GMC audit data, live account flags, fix plans, and appeal generation. No copy-pasting. No switching tabs. Your AI just handles it.

🤖Claude Desktop⌨️Cursor🏄Windsurf🔷Cline (VS Code)🔄Continue.dev💻Claude Code⚡Zed Editor🪿Goose🧙Witsy🔌Any stdio MCP

Get your API key Setup guide npm package

Free to set up · Works with your existing GMC Guard plan · Not affiliated with Google

The gap Google left open

There is no official Google MCP server. We built it.

AI agents can connect to Slack, GitHub, Linear, Notion, Stripe — but not to Google Merchant Center. There’s no way for Claude or any other AI to read your GMC account status, see what products got flagged, or understand why you were suspended.

GMC Guard fills that gap. Our MCP server is the bridge between the AI tools you already use and the Google Merchant Center data you need.

Live GMC account status + product flags via the Merchant API
Full audit engine — the same AI that powers gmcguard.ai
Appeal generation informed by what Google actually sees
Weekly monitoring your AI can check and report on
Works with Claude, Cursor, Windsurf, Cline, Goose, Zed and more

Claude — GMC Guard MCP connected

8 tools active

9 tools available

What your AI agent can do

Every tool is described so your AI knows exactly when and why to use it. Ask a question — the agent picks the right tools automatically.

list_audits

See all audited stores with trust scores, risk levels, and monitoring status.

“What stores have I audited and what are their scores?”

run_audit

core

Trigger a full misrepresentation audit on any domain. Returns trust score + prioritised issue list.

“Audit mystore.com for GMC misrepresentation issues.”

get_audit

Get every issue from an audit with severity, category, and paste-ready fix instructions.

“Show me all the critical issues from my last audit.”

get_fix_plan

core

Prioritised fix plan grouped by category — policy, identity, product claims, feed. Exact fix copy for each step.

“What do I need to fix before I can appeal?”

generate_appeal

core

Generate a ready-to-submit Google re-review appeal based on what was found and fixed.

“Write my Google re-review appeal based on the fixes I've made.”

check_gmc_account

Read live GMC account status, connected accounts, and product-level flags direct from the Merchant API.

“What does my Google Merchant Center account look like right now?”

get_monitoring

Score history and monitoring status for all your stores. See trends over time.

“Has my trust score improved since I made changes last week?”

set_monitoring

Enable or disable weekly monitoring on any domain. Get alerted when scores drop.

“Start monitoring mystore.com every week.”

create_checkout

new

Generate a secure checkout link so the user can upgrade their plan — pre-filled with their email and details. AI never handles payment data.

“I want to upgrade to Pro so I can run more audits.”

Who uses it

Three types of GMC problems, one MCP connection

Suspended

Merchant fixing a suspension

Open Claude Desktop, type “my GMC got suspended, audit mystore.com.” Your AI runs the audit, shows every issue with exact fix copy, then writes the appeal when you’re done. No dashboard, no copy-pasting.

Audit → Fix plan → Appeal in one conversation
AI explains each issue in plain language
Re-check score after fixing to confirm progress

Agency

Agency running client audits

Build an internal AI workflow that audits every client store, flags score drops, and prepares a weekly summary report — all automated via MCP tool calls.

Batch-audit all client domains
AI writes the client-facing report
Monitor across all stores simultaneously

Developer

Developer building GMC tooling

Integrate GMC Guard into your own AI agent or automation. Use the REST API directly or wire the MCP server into any workflow that supports the protocol.

RESTful v1 API with Bearer auth
stdio MCP transport (any client)
npm package — one npx command

No setup required

Just paste this into your AI

Not ready to configure MCP? Copy this prompt and paste it directly into Claude.ai, ChatGPT, or any AI chat. Add your API key to unlock live tool calls.

Agent starter prompt

You are connected to GMC Guard — a Google Merchant Center misrepresentation audit tool.

You have access to these tools (requires GMCGUARD_API_KEY):
- list_audits — see all audited stores and scores
- run_audit(domain) — audit any store for GMC suspension risk
- get_audit(audit_id) — full results with issues and fix instructions
- get_fix_plan(audit_id) — prioritised fix plan by category
- generate_appeal(audit_id) — write a Google re-review appeal
- check_gmc_account — live GMC account status and flags
- get_monitoring — score history and monitoring status
- set_monitoring(domain, enabled) — enable/disable weekly monitoring
- create_checkout(plan) — generate a secure upgrade link for the user

Rules:
- Always get real data before giving advice
- Never recommend appealing while criticals > 0
- Appeal is ready when: criticals = 0 AND trust score ≥ 70
- Monitoring requires Pro, Watch, or Agency plan

Start by asking what domain they want to check, or call list_audits to show history.

For full live tool access, generate an API key and follow the MCP setup below.

Connect in 3 steps

Works with every major AI client

Claude Desktop, Cursor, Windsurf, Cline, Continue.dev, Goose, Zed, Witsy and more. All use the same npm package — just change the config format.

Get your API key

Go to your GMC Guard profile → Generate key in the API Key / MCP Access section. Copy it — it’s only shown once.

Add the MCP server to your AI client

Select your client below. No installation needed — npx downloads and runs the server automatically.

Edit ~/.claude/claude_desktop_config.json

{
  "mcpServers": {
    "gmcguard": {
      "command": "npx",
      "args": ["-y", "gmcguard-mcp-server"],
      "env": {
        "GMCGUARD_API_KEY": "gmcg_your_key_here"
      }
    }
  }
}

Restart and start asking

Restart your AI client. The GMC Guard tools appear automatically. Try:

"My Google Merchant Center is suspended. Audit mystore.com and tell me what to fix."

"What's my current trust score and are there any critical issues?"

"I've made all the fixes — write my appeal letter."

"Start weekly monitoring for mystore.com."

Direct API access

REST API for custom integrations

Don’t need the full MCP server? Use the v1 REST API directly.

GET/api/v1/auditsList recent audits

POST/api/v1/auditsTrigger new audit

GET/api/v1/audits/:idFull results + issues

GET/api/v1/gmcLive GMC account status

GET/api/v1/monitoringScore history

PATCH/api/v1/monitoringEnable/disable monitoring

POST/api/v1/checkoutGenerate checkout URL

curl https://gmcguard.ai/api/v1/audits \
  -H "Authorization: Bearer gmcg_your_key"

Plan requirements

What you need

API key + MCP connectionAll plans

list_audits / get_auditAll plans

check_gmc_accountAll plans

run_audit (standard)All plans (credits)

run_audit (advanced)Pro, Agency

get_fix_planAll plans

generate_appealPro, Agency

get_monitoring / set_monitoringPro, Watch, Agency

create_checkoutAll plans

View plans →

Ready to connect

Let your AI handle the GMC work

Get your API key, paste one config block into your AI client, and your assistant can audit, fix, and monitor Google Merchant Center accounts from inside any conversation.

Generate API key See setup guide

Free account · No credit card · API key available immediately after sign-up