Google Merchant Center MCP

Shows whether this server is connected to Google: token presence and source (env variables or a stored in-chat login), expiry, the Google account email, granted vs missing OAuth scopes, where the credentials file lives and where the OAuth client comes from. Makes no network calls and never returns the token itself. Call it first when other tools report the server is not connected.

Takes no parameters.

Step-by-step checklist for connecting this server to Google: creating a Google Cloud project and a Desktop-app OAuth client, publishing the consent screen (mandatory — Testing-mode refresh tokens die after 7 days), downloading the client JSON and handing its PATH to set_client. Works without any credentials; the checklist shortens to 'enable the API + log in' when an OAuth client is already configured (one client serves the whole mcp-google-* line). Never asks the user to paste secrets into the chat.

Takes no parameters.

Saves the OAuth client credentials from the JSON file downloaded from Google Cloud Console ('Download JSON' on a Desktop-app client). Pass the file PATH — the secret must never be pasted into the chat. The client is stored once in the shared ~/.config/mcp-google-auth/client.json (owner-only) and reused by every mcp-google-* server; tokens stay per-server. After this, call start_login.

First step of connecting from the chat, without editing config files or restarting the client. Returns authorizeUrl — show it to the user as a clickable link and ask them to open it in the browser ON THIS MACHINE, pick the Google account and approve access. A one-shot listener on 127.0.0.1 catches Google's redirect; the code is exchanged locally and never passes through the chat. Does not open the browser itself. The attempt lives 10 minutes; when the browser shows the success page, call finish_login.

Takes no parameters.

Second step: waits up to 20 s for the browser consent to finish, saves the tokens to an owner-only file and verifies the login with a read-only identity call, returning the account email and the granted scopes. After success every tool works immediately — no client restart. If the user granted only part of the requested permissions, the login is still saved and missingScopes lists what will not work. Logging in under a different Google account replaces the previous login (its refresh token is revoked best-effort) and the response carries previousAccountEmail so the change never goes unnoticed.

Takes no parameters.

Revokes the stored token at Google (oauth2.googleapis.com/revoke) and deletes the local credentials file. Tokens supplied via env variables are NOT touched — remove them from the MCP client config manually; envTokenStillSet in the response says whether any are still in effect.

Takes no parameters.

Lists the Merchant Center accounts the authenticated user can access. Returns accounts[] (name, accountId, accountName, languageCode, timeZone, adultContent, testAccount) and nextPageToken. Use it first to discover the account ID the other tools need (or set GOOGLE_MERCHANTS_ACCOUNT_ID once). Optional filter uses the account filter syntax, e.g. accountName = "*store*".

Returns a single Merchant Center account: name (accounts/{id}), accountId, accountName, languageCode, timeZone, adultContent and testAccount. Useful to verify the configured account or inspect a sub-account.

Returns the store homepage of an account: uri and claimed (whether the homepage is verified and claimed by the merchant — a prerequisite for serving offers). An unclaimed homepage is a common reason for account-level problems; claiming/unclaiming is not exposed as a tool (use raw_request POST accounts/v1/accounts/{a}/homepage:claim if you really need it).

Returns the account-level shipping settings: services[] (delivery countries, delivery times, rate tables and carrier rates), warehouses[] and an etag. Read-only by design: the API's only write is shippingSettings:insert, a FULL REPLACE of every service — too dangerous for a tool; use raw_request if you really need it.

Lists the processed products of an account, as shown in Merchant Center. Each product has name (accounts/{a}/products/{contentLanguage~feedLabel~offerId} — NO channel segment in v1), offerId, contentLanguage, feedLabel, dataSource, productAttributes (title, price, availability, ...), productStatus with itemLevelIssues, and base64EncodedName (use it when offerId contains URL-hostile characters like '/'). The list has no server-side filter — filter via search_reports on product_view. Recently inserted products appear only after async processing (minutes).

Returns one processed product including productStatus.itemLevelIssues (code, severity, resolution, description) — the place to see why a product is disapproved. Identify the product either with `product` ("contentLanguage~feedLabel~offerId", e.g. "en~US~sku123", or the base64url base64EncodedName; legacy local products use a local~ prefix) or with the three components content_language + feed_label + offer_id. A product inserted moments ago may 404 until async processing finishes (minutes).

Uploads (upserts) a product into an API data source: an existing input with the same contentLanguage~feedLabel~offerId in that data source is fully replaced. Requires data_source (an API-type source — create one with create_data_source or in Merchant Center; file feeds cannot be written). Inserting with a different data source MOVES the product to it. Returns the ProductInput (name, product = the future processed name, base64EncodedProduct). Processing is async: the processed product shows up in get_product/list_products after several minutes, and data-quality problems surface later in productStatus.itemLevelIssues, not as API errors. Prices go in product_attributes as {"price": {"amountMicros": "9990000", "currencyCode": "USD"}} (1 unit = 1,000,000 micros).

Sparse-updates an existing product input — the cheap way to change price or availability without re-sending the whole product. data_source must be the source holding the input. update_mask is a comma-separated list of attribute paths (e.g. "productAttributes.price,productAttributes.availability"); when omitted, all populated fields of the request are applied. Every path listed in update_mask MUST carry a value in this request — a masked path with no value ERASES that attribute (the tool rejects such requests locally; to clear an attribute intentionally use raw_request). Returns the updated ProductInput; the processed product refreshes after async processing (minutes). To create a product or replace it wholesale use insert_product_input.

Deletes a product input from a specific data source (data_source is required — the same product can exist in several sources, and only the targeted input is removed). product_input is "contentLanguage~feedLabel~offerId" or the base64url name. Returns an empty object on success; the processed product disappears after async processing (minutes).

Lists the data sources of an account. Each has name (accounts/{a}/dataSources/{id}), dataSourceId, displayName, input (API | FILE | UI | AUTOFEED), exactly one type object (primaryProductDataSource, supplementalProductDataSource, promotionDataSource, ...) and fileInput for file feeds. Use it to find the API-type data source that insert_product_input / insert_promotion require as data_source.

Returns one data source by its numeric ID (or full resource name): type, input (API/FILE/UI/AUTOFEED), feed configuration and fetch settings. Check `input` before calling fetch_data_source — only file-based feeds with fetch settings can be re-fetched.

Creates an API (generic) data source — the target that insert_product_input / update_product_input / insert_promotion need as data_source. Only API sources can be created through the API (file, UI and autofeed sources are set up in Merchant Center). For product sources content_language and feed_label must be both set or both omitted; countries applies to primary sources only. A promotions source requires target_country and content_language. Returns the created DataSource with its dataSourceId.

Triggers an immediate re-fetch (re-read) of a file-based feed outside its regular schedule. Works only for data sources with a file input and fetch settings (scheduled fetch or Google Sheets); calling it on an API-type source is an error. Returns an empty object on success — the fetch itself runs asynchronously on Google's side.

Creates or updates a promotion. Unlike product writes, the data source travels in the request BODY (the server assembles it). The promotion object requires promotionId, contentLanguage (ISO 639-1), targetCountry (CLDR, e.g. "US") and redemptionChannel (array with ONLINE and/or IN_STORE — at least one). Optional `attributes` carry longTitle, couponValueType, offerType, genericRedemptionCode, promotionEffectiveTimePeriod {startTime, endTime}, productApplicability, moneyOffAmount, percentOff, etc. Returns the Promotion incl. promotionStatus; approval happens asynchronously (check get_promotion later).

Lists the promotions of an account: promotions[] (name, promotionId, contentLanguage, targetCountry, redemptionChannel, attributes, promotionStatus with destination statuses and itemLevelIssues) and nextPageToken.

Returns one promotion including promotionStatus (per-destination approval and itemLevelIssues) — the place to check whether a freshly inserted promotion was approved.

Runs a Merchant Center Query Language (MCQL) query via reports:search. Tables: product_view, product_performance_view, price_competitiveness_product_view, price_insights_product_view, non_product_performance_view, best_sellers_product_cluster_view, best_sellers_brand_view, competitive_visibility_top_merchant_view, competitive_visibility_competitor_view, competitive_visibility_benchmark_view. Rules: field names are snake_case in the query but camelCase in the JSON response; no SELECT *; performance views require a WHERE date range, e.g. SELECT offer_id, clicks, impressions FROM product_performance_view WHERE date BETWEEN '2026-07-01' AND '2026-07-31' ORDER BY clicks DESC. price_* views require the Market Insights opt-in. Each result row has exactly one populated view object. This is also the way to FILTER products (product_view) — list_products has no filter.

Convenience wrapper over a canned MCQL query on price_competitiveness_product_view: for each product, your price vs the market benchmark_price (aggregated from comparable offers across merchants) with report_country_code. A product priced above the benchmark is losing clicks to cheaper rivals. Requires the account to be opted into Market Insights (free, in Merchant Center settings) — otherwise rows are empty. Price amounts are micros (1,000,000 = 1 unit) and may arrive as strings (int64). Optional country narrows to one report country.

Convenience wrapper over a canned MCQL query on price_insights_product_view: Google's suggested_price per product with the predicted change in impressions, clicks and conversions if you adopt it (predicted_*_change_fraction, e.g. 0.05 = +5%), plus an overall `effectiveness` bucket (LOW/MEDIUM/HIGH). Requires the Market Insights opt-in — otherwise rows are empty. Price amounts are micros and may arrive as strings (int64).

Lists aggregate product statuses per reporting context and country (issueresolution sub-API): stats {active, pending, disapproved, expiring counts} plus itemLevelIssues[] with how many products each issue affects — the fastest way to see what is wrong with a feed at a glance. Works only for sub-accounts and standalone accounts, NOT for advanced (parent) accounts. The filter supports only reporting_context and country, e.g. reporting_context = "SHOPPING_ADS" AND country = "US". For a single product's issues use get_product (productStatus.itemLevelIssues).

Shows the account's Merchant API usage vs limits per method group (quota sub-API): quotaGroups[] with name, quotaUsage, quotaLimit (per day), quotaMinuteLimit and methodDetails[] listing the methods in each group. Daily counters reset at 12:00 UTC — midday, not midnight. Use it to diagnose HTTP 429 RESOURCE_EXHAUSTED errors and to see how much headroom is left.

Escape hatch to call any Merchant API v1 path directly, for endpoints without a dedicated tool (e.g. "accounts/v1/accounts/123/issues" or the one-time "accounts/v1/accounts/123/developerRegistration:registerGcp"). The path must include the sub-API prefix (accounts/v1, products/v1, datasources/v1, promotions/v1, reports/v1, issueresolution/v1, quota/v1, inventories/v1, ...). `query` adds URL query parameters (e.g. dataSource for productInputs writes, updateMask for PATCH); `body` is sent as JSON. Can create, modify and delete data — use the dedicated tools when one exists.