Product Feed Doctor:find what is wrong withyour shopping feedbefore Google, Metaor your AI assistant do

Google Shopping / Merchant Center RSS 2.0 (the g: namespace format, with or without an atom:link tag), a Facebook/Meta catalog CSV, a Shopify /products.json export, a WooCommerce Store API JSON response, a generic XML feed built from repeating <item> elements with different tag names, and a plain CSV with a header row. The format is auto-detected from the content you paste, upload, or fetch; you never have to tell it which one you're using.

No. Parsing and every check run entirely in your browser, in feed-doctor.js. If you paste a feed URL and click Fetch, your browser fetches that URL directly; the response never passes through any ARLing server. Pasted text, an uploaded file, and the report are never sent anywhere. Read feed-doctor.js directly, or check your browser's Network tab, to verify.

Most shop platforms and CDNs do not send the Access-Control-Allow-Origin header on their feed URLs, since those files are meant to be downloaded by Google's or Meta's servers, not read by JavaScript running on a different website. When that happens, the browser blocks the response before this page ever sees it: that is normal CORS behavior, not a bug in this tool. Paste the feed content instead, or upload the downloaded file; both skip the browser fetch entirely and analyze the exact same content.

The Merchant Center product data specification treats id, title, description, link, image_link, price a availability as required for essentially every item, and condition as required once an item is not new (used, refurbished, and similar). This tool flags all of those as errors when missing or blank, and additionally checks brand, gtin, mpn, item_group_id a shipping, which the specification calls for in specific situations (a branded, identifiable product; variants that need to be told apart; and so on).

gtin is checked against the standard GS1 mod-10 check-digit algorithm used by GTIN-8, GTIN-12 (UPC-A), GTIN-13 (EAN-13) and GTIN-14: non-digit characters are stripped, the length must be one of those four, and the last digit must equal (10 minus the mod-10 remainder of the weighted sum of the preceding digits, alternating weights 3 and 1 from the right) mod 10. A gtin that is the wrong length, or has a transposed or mistyped digit, fails this check and is flagged with the exact value that was read.

It is a simple, documented heuristic, not a number Google, Meta or any platform publishes: start at 100, subtract 3 points per error instance, 1 point per warning instance and 0.2 points per info instance found across the whole feed, then floor at 0 and round to the nearest integer. It is meant to give a quick sense of how much is wrong and how severe it is, not to predict feed approval; read the rule-by-rule report for the specifics that matter.

It is an opt-in add-on: give it your feed URL and e-mail after an analysis, and a server checks the same 31 rules on a schedule (weekly on the free plan, daily on Pro) and e-mails you when something got worse. The checker above stays entirely local to your browser; the monitor is the only part of Feed Doctor that fetches your feed on our server, and it only fetches the URL you gave it.

The free plan covers one feed per e-mail address, a weekly check, and an alert when a new error rule appears or the score drops by 5 points or more. Pro is 9 EUR per month per feed: a daily check, an alert on every change, and 90 days of score history on your monitor page. Unsubscribe with one click from any alert e-mail or the monitor page.

Your e-mail, the feed URL, and the result of each check: the score, the counts of errors/warnings/infos, and which rule ids triggered. It never stores the feed content itself, so your product titles, prices and descriptions never sit on our server.