mcp-google-merchants

mcp-google-merchants

Enables AI assistants to manage Google Merchant Center accounts via the Merchant API, including product feeds, promotions, data sources, and MCQL reports with read and write operations.

Category
Visit Server

README

Google Merchant Center MCP

npm CI Glama License: MIT

MCP server for Google Merchant Center via the Merchant API v1: manage your product feed, promotions and data sources, run MCQL reports, check price competitiveness and product issues — from Claude, Cursor, Codex and other AI clients in natural language.

Unlike read-only integrations, this server authenticates with your own OAuth credentials and supports write operations: upload and delete products, insert promotions and trigger feed re-fetches — with destructive tools clearly annotated so MCP hosts can gate them.

Quick start

  1. Get OAuth credentials for the Merchant API.

  2. Add the server — for example in Claude Code (other clients):

    claude mcp add google-merchants \
      -e GOOGLE_MERCHANTS_CLIENT_ID=your_client_id \
      -e GOOGLE_MERCHANTS_CLIENT_SECRET=your_client_secret \
      -e GOOGLE_MERCHANTS_REFRESH_TOKEN=your_refresh_token \
      -e GOOGLE_MERCHANTS_ACCOUNT_ID=your_merchant_id \
      -- npx -y mcp-google-merchants@latest
    
  3. Ask the assistant: "Which of my products are disapproved, and why?"

What it can do

Tool Description
list_accounts Merchant Center accounts you can access (with optional filter).
get_account One account's settings (name, language, time zone, ...).
get_homepage The store homepage and whether it is claimed.
get_shipping_settings Account-level shipping services and warehouses.
list_products Processed products as shown in Merchant Center, incl. statuses.
get_product One product with itemLevelIssues — why it is disapproved.
insert_product_input Upload (upsert) a product into an API data source.
update_product_input Sparse-update a product (price, availability, ...).
delete_product_input Delete a product input from a data source.
list_data_sources Feeds/data sources of the account (API, file, UI, autofeed).
get_data_source One data source with its feed and fetch configuration.
create_data_source Create an API data source for product/promotion writes.
fetch_data_source Trigger an immediate re-fetch of a file feed.
insert_promotion Create or update a promotion.
list_promotions Promotions with their approval statuses.
get_promotion One promotion incl. per-destination status.
search_reports Run any MCQL query (reports:search).
price_competitiveness Your prices vs market benchmarks (canned MCQL).
price_insights Google's suggested prices + predicted impact (canned MCQL).
list_product_issues Aggregated product issues per reporting context/country.
list_method_quotas API usage vs quota limits per method group.
raw_request Escape hatch to any Merchant API v1 path (SSRF-guarded).

Resilience: retries with backoff on 429 and on 5xx/network errors for reads (writes are never replayed), Retry-After support, request timeouts, automatic access-token refresh.

Example prompts

  • "List my Merchant Center products that are out of stock"
  • "Why is product sku-123 disapproved in Shopping ads?"
  • "Upload a test product 'Blue Widget' for $9.99 to my API feed"
  • "Which of my products are priced above the market benchmark in the US?"
  • "Show clicks and impressions per product for July"

MCQL examples

search_reports accepts raw Merchant Center Query Language. Field names are snake_case in queries and camelCase in responses; SELECT * is not supported; performance views require a date range.

-- Filter products (list_products has no filter — this is the way)
SELECT offer_id, title, price, aggregated_reporting_context_status
FROM product_view
WHERE aggregated_reporting_context_status = 'NOT_ELIGIBLE_OR_DISAPPROVED'
-- Performance over a date range
SELECT offer_id, title, clicks, impressions, click_through_rate
FROM product_performance_view
WHERE date BETWEEN '2026-07-01' AND '2026-07-31'
ORDER BY clicks DESC
-- Price competitiveness (requires the free Market Insights opt-in)
SELECT offer_id, title, price, benchmark_price
FROM price_competitiveness_product_view
WHERE report_country_code = 'US'

API access

The server talks to the Merchant API v1 (merchantapi.googleapis.com) — the successor of the Content API for Shopping (sunset in August 2026). Auth is standard Google OAuth 2.0 with the scope https://www.googleapis.com/auth/content; the server exchanges your refresh token for access tokens automatically.

One-time registration required. Before any Merchant API call works, your Google Cloud project must be registered with the Merchant Center account once (needs Admin access):

raw_request POST accounts/v1/accounts/{account}/developerRegistration:registerGcp
body: {"developerEmail": "you@example.com"}

You can do it right from the assistant with the raw_request tool, or with any HTTP client. Until then every call fails with a permission error.

Installation

<details open> <summary><b>Claude Code</b></summary>

claude mcp add google-merchants \
  -e GOOGLE_MERCHANTS_CLIENT_ID=your_client_id \
  -e GOOGLE_MERCHANTS_CLIENT_SECRET=your_client_secret \
  -e GOOGLE_MERCHANTS_REFRESH_TOKEN=your_refresh_token \
  -e GOOGLE_MERCHANTS_ACCOUNT_ID=your_merchant_id \
  -- npx -y mcp-google-merchants@latest

</details>

<details> <summary><b>Claude Desktop</b></summary>

claude_desktop_config.json — macOS ~/Library/Application Support/Claude/, Windows %APPDATA%\Claude\

{
  "mcpServers": {
    "google-merchants": {
      "command": "npx",
      "args": ["-y", "mcp-google-merchants@latest"],
      "env": {
        "GOOGLE_MERCHANTS_CLIENT_ID": "your_client_id",
        "GOOGLE_MERCHANTS_CLIENT_SECRET": "your_client_secret",
        "GOOGLE_MERCHANTS_REFRESH_TOKEN": "your_refresh_token",
        "GOOGLE_MERCHANTS_ACCOUNT_ID": "your_merchant_id"
      }
    }
  }
}

</details>

<details> <summary><b>Cursor</b></summary>

~/.cursor/mcp.json (or .cursor/mcp.json in a project)

{
  "mcpServers": {
    "google-merchants": {
      "command": "npx",
      "args": ["-y", "mcp-google-merchants@latest"],
      "env": {
        "GOOGLE_MERCHANTS_CLIENT_ID": "your_client_id",
        "GOOGLE_MERCHANTS_CLIENT_SECRET": "your_client_secret",
        "GOOGLE_MERCHANTS_REFRESH_TOKEN": "your_refresh_token",
        "GOOGLE_MERCHANTS_ACCOUNT_ID": "your_merchant_id"
      }
    }
  }
}

</details>

<details> <summary><b>VS Code</b></summary>

.vscode/mcp.json — note the servers key (not mcpServers)

{
  "servers": {
    "google-merchants": {
      "type": "stdio",
      "command": "npx",
      "args": ["-y", "mcp-google-merchants@latest"],
      "env": {
        "GOOGLE_MERCHANTS_CLIENT_ID": "your_client_id",
        "GOOGLE_MERCHANTS_CLIENT_SECRET": "your_client_secret",
        "GOOGLE_MERCHANTS_REFRESH_TOKEN": "your_refresh_token",
        "GOOGLE_MERCHANTS_ACCOUNT_ID": "your_merchant_id"
      }
    }
  }
}

</details>

Getting access

  1. Create (or pick) a Google Cloud project at console.cloud.google.com and enable the Merchant API (APIs & Services → Library → "Merchant API" → Enable).

  2. Configure the OAuth consent screen (APIs & Services → OAuth consent screen): External, fill in the app name and your email, and add yourself as a test user (Testing mode is fine for personal use).

  3. Create an OAuth client (APIs & Services → Credentials → Create credentials → OAuth client ID → Desktop app). Save the client ID and client secret.

  4. Mint a refresh token for the scope https://www.googleapis.com/auth/content. The quickest way is the OAuth 2.0 Playground:

    • click the gear icon → check Use your own OAuth credentials → paste your client ID/secret;
    • in Step 1 enter the scope https://www.googleapis.com/auth/content and authorize with the Google account that has access to your Merchant Center;
    • in Step 2 click Exchange authorization code for tokens and copy the refresh token.

    (Any other flow works too — the server only needs the resulting refresh token. For quick experiments you can instead pass a short-lived access token as GOOGLE_MERCHANTS_ACCESS_TOKEN, e.g. from gcloud auth print-access-token.)

  5. Find your Merchant Center ID — the number in the top-right corner of merchants.google.com — and put it in GOOGLE_MERCHANTS_ACCOUNT_ID (or pass account per tool call, or discover it with list_accounts).

  6. Register your GCP project with the Merchant Center account (one-time, Admin access required) — see API access.

⚠️ Credentials are stored in plain text in your MCP client config — treat them like passwords. The refresh token grants full read/write access to your Merchant Center.

Configuration

Variable Required Default Description
GOOGLE_MERCHANTS_CLIENT_ID yes* OAuth 2.0 client ID.
GOOGLE_MERCHANTS_CLIENT_SECRET yes* OAuth 2.0 client secret.
GOOGLE_MERCHANTS_REFRESH_TOKEN yes* OAuth refresh token (scope .../auth/content).
GOOGLE_MERCHANTS_ACCESS_TOKEN yes* Pre-minted access token (~1h) — alternative to the three above.
GOOGLE_MERCHANTS_ACCOUNT_ID no Default Merchant Center account ID; tools can override per call.
GOOGLE_MERCHANTS_API_BASE no https://merchantapi.googleapis.com API root override.
GOOGLE_MERCHANTS_TOKEN_URL no https://oauth2.googleapis.com/token OAuth token endpoint override.
GOOGLE_MERCHANTS_TIMEOUT_MS no 60000 Per-request timeout, ms.
GOOGLE_MERCHANTS_MAX_RETRIES no 3 Retries on transient errors.

* Either the client ID + secret + refresh token trio, or a bare access token.

Requirements

  • Node.js 20+ (runs via npx, no separate install needed).
  • A Merchant Center account and a registered Google Cloud project — see Getting access.

Limitations

  • Product writes are asynchronous. After insert_product_input / delete_product_input the processed product updates within minutes — an immediate get_product may 404 or show stale data. Data-quality problems appear later in productStatus.itemLevelIssues, not as API errors.
  • price_competitiveness / price_insights return rows only for accounts opted into Market Insights (free, in Merchant Center settings).
  • list_product_issues works for sub-accounts and standalone accounts only, not for advanced (parent) accounts.
  • Quotas are per-account with daily counters resetting at 12:00 UTC (midday). Check your current usage with list_method_quotas.

Documentation

Support

Questions, ideas and contributions — Telegram: @gistrec or GitHub issues.

License

MIT — see LICENSE.

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured