BEFORE THE IMPORT
A cleaner catalogue.
Without guessing.
Match unfamiliar product columns with a local open model. Review the mapping, then catch catalogue errors before they reach your shop.
01 Add a CSV02 Review columns03 Export the audit
01 / Your product CSV
English headersFiles stay in this browser. The model is served locally; there is no hosted AI endpoint.
Ready for a CSV.
What gets checked?
SKU, title, finite nonnegative price, supported currency, repeated SKUs and row widths. This prototype is not a Shopify compatibility certificate or a tax/pricing check.
02 / Review the mapping
Waiting for CSVSuggestions are provisional. Weak, ambiguous or conflicting matches stay unselected.
Your column suggestions appear here.
The file is never rewritten. Original source line numbers are kept in the report.
03 / Audit report
Review the mapping to see the audit.
Prototype evidence
Frozen synthetic header fixtures compare the local model with an exact-alias baseline. This small test is not a customer accuracy benchmark.
Validation has not run in this browser.
Local model and request evidence
Model not loaded.