How the Merchant Feed Validator Works
AnyTool validates product feeds entirely in the browser. The CSV/TSV feed is parsed with PapaParse, headers are resolved through an alias map (image url/img_link → image_link, sku/product_id → id, …), and each product is checked against the selected spec — Google Merchant Center, Meta/Facebook or Microsoft. Checks include required fields (reported once as a structural "column absent" vs per-row "empty value"), title and description length, price format (numeric with optional currency, parsed locale-aware), image and product URLs (http/https), availability allow-lists, duplicate IDs across rows, and an all-caps/excessive-punctuation title heuristic based on a real uppercase ratio. Issues are tiered error/warning/info into a weighted 0–100 health score, with per-field counts and duplicate-ID and missing-image stats, exportable as CSV or JSON. Nothing is uploaded.
- Runs client-side — the feed is never uploaded
- Google / Meta / Microsoft spec rule packs + header aliases
- Required/length/price/URL/availability/duplicate-ID checks
- Real uppercase-ratio title check (not a literal string match)
- Weighted 0–100 score; CSV/JSON issue export
- No CDN, no backend, no product cap or signup
