Catalog QA Desk in ChatGPT
Product catalog checks. The MCP server is https://catalogqa.openkrill.app/mcp. No key. The steps below are only for ChatGPT. A call to csv_import_preflight was checked against that server on 2026-10-01.
What this server answers
Check an online store's product catalog before you import it or send it to a shopping feed.
Ask "will this Shopify product CSV import cleanly?", "what would Google Merchant Center reject in my catalog?", "which products have no image alt text or a copied description?" or "write me a sheet that fixes the empty alt text".
Catalog QA Desk checks a Shopify product CSV against the import rules (required columns, handles, options, duplicate variants, prices, statuses, image URLs, barcodes mangled by a spreadsheet) and returns the rows that block the import and the ones that import wrongly, each with the fix.
It checks products for Google Merchant Center feed rules: missing or invalid barcodes (GTIN length and check digit), titles that are too long, promotional or in capitals, missing descriptions, links in descriptions and missing images.
It audits listing text: images without alt text, descriptions copied between products, thin descriptions and title problems; and it builds an advisory CSV of corrected titles and alt text for you to review and import yourself.
You can pass the CSV text (up to 2 MB), or a store address: the store's public product feed (/products.json, at most 5 pages of 250 products) is read once, after its robots.txt, only when you ask. Large stores often block that feed, and then a CSV is needed.
It does not use the Shopify admin API, does not log in to any store, changes nothing in a store and does not submit anything to Google. Nothing you send or read is stored, no catalog is kept or indexed, and there is no account; only a daily count of calls per tool is recorded.
What it can do
- Checks a Shopify product CSV before import and lists blocking rows and warnings, each with the fix
- Finds products Merchant Center would reject: bad or missing GTINs, long or promotional titles, empty descriptions
- Finds images without alt text, copied or thin descriptions and title problems
- Builds a CSV of corrected titles and alt text to review and import; changes nothing in the store
- Reads a store's public product feed once on request, after robots.txt, instead of a CSV
Tools
csv_import_preflight, Check a product CSV before import. Check the text of a Shopify product CSV against the import rules before it is imported. Use it for "will this CSV import?", "what is wrong with my product file?". Pass csv exactly as exported or edited, header row included, up to 2 MB. Returns blocking findings (Shopify rejects them or the import comes out wrong) and warnings, up to 200 each, every one with row, handle, rule and the fix, plus totals. Nothing is stored: the file or feed is used for this answer and dropped.merchant_feed_check, Check products for Merchant Center rules. Check products against Google Merchant Center feed rules: missing or invalid GTIN barcodes, long, promotional or all-capital titles, missing descriptions, links, images and brand. Use it for "what would Merchant Center reject?". Pass csv (a Shopify product CSV) or domain (a store address), not both. Returns a summary (products, no_gtin, bad_gtin, long_title, no_description) and up to 200 issues, one per product and rule, with detail. Domain mode reads the store's public product feed once after its robots.txt; a store that blocks it needs a CSV. The public feed has no barcodes, so in domain mode no_gtin and bad_gtin are null and GTINs are not checked; use a CSV for that. Nothing is stored: the file or feed is used for this answer and dropped.listing_text_audit, Audit product listing text. Audit the text of product listings: images without alt text, descriptions copied between products, descriptions that are too short and title problems. Use it for "which products have no alt text?", "are any descriptions duplicated?". Pass csv or domain, not both, and optionally min_words (default 20). Returns the four lists (each up to 100 entries) with totals before the cut. Domain mode reads the store's public product feed once after its robots.txt. Nothing is stored: the file or feed is used for this answer and dropped.listing_fix_sheet, Build a CSV of listing text fixes. Build a CSV of corrected titles and image alt text from a catalog, for the user to review and import. Use it for "write the alt text for my images", "fix the shouting titles". Pass csv or domain, not both; optional fixes (alt_text, title_whitespace, title_caps) and alt_text_template with {title}, {vendor}, {type}, {position}. Returns csv with only the rows that change (URL handle, Title, Product image URL, Image position, Image alt text), up to 200 rows, and counts. Advisory: nothing is changed in any store. Nothing is stored: the file or feed is used for this answer and dropped.
Add Catalog QA Desk in ChatGPT
ChatGPT does not read a local config file. A custom MCP server is added in the product, and only after developer mode is on. These steps were checked against OpenAI's plugin connect guide and the ChatGPT MCP guide on 2026-10-01. Developer mode can be hidden by a workspace policy, so a Team or Enterprise admin may have to allow it first.
- Open ChatGPT settings, then Security and login, and turn on Developer mode.
- Open ChatGPT plugins and select the plus button.
- Enter a name such as "Catalog QA Desk" and a one-line description of what you want it to answer.
- Under Connection, choose the public endpoint and paste
https://catalogqa.openkrill.app/mcp. Include the/mcppath. Do not add a tunnel id: this server is already public. - Create the connection and review the tools ChatGPT lists. You should see
csv_import_preflight,merchant_feed_check,listing_text_audit,listing_fix_sheet. The feedback tools on the same server are optional and not required for the lookup. - Start a new conversation, add the connection from the tools menu, and ask: Will this Shopify product CSV import cleanly?
No API key and no OAuth client id are required. The server answers without a login. Calls are limited per network and per day, and a limit comes back as a tool error rather than a partial answer. If the tool list is stale after a server change, open the connection and select Refresh, then start a new conversation.
The ChatGPT desktop app, Codex CLI and IDE extension share a different config. In the desktop app, open Settings, then MCP servers, choose Add server, select Streamable HTTP, and paste the same URL. In ~/.codex/config.toml the equivalent is a table named mcp_servers.catalog-qa with url = "https://catalogqa.openkrill.app/mcp". ChatGPT on the web does not read that file. A published directory listing is also a different path: this page is the developer-mode connection, not a claim that Catalog QA Desk is already in the ChatGPT directory.
Treat the server as untrusted until you have read a few answers. A remote MCP server sees the arguments the model sends. Catalog QA Desk is read-only and documents what it sends onward on the privacy page.
A call checked on 2026-10-01
A two-column Shopify product CSV is enough for the import preflight to return row findings. The request below was posted to https://catalogqa.openkrill.app/mcp as tools/call. HTTP 200. Source: the Catalog QA Desk server, read 2026-10-01.
{
"method": "tools/call",
"params": {
"name": "csv_import_preflight",
"arguments": {
"csv": "Handle,Title,Variant Price\nmug,Blue mug,12\n"
}
}
}- warnings. 1 returned
Repeat the call yourself if you need a newer reading. Cached answers expire. A rate limit is not a result: wait and try again. Nothing in the call is a ranking, a filing, or advice.