All products

Localise a Persian storefront with AI

Review mixed-direction product names, checkout spacing and display formats while keeping catalogue identifiers and stored amounts consistent.

Localise a Persian storefront with AI using a realistic basket as your test case. Put a Latin model code beside the Persian product name and add a discount. The rtl-and-persian-localization skill helps you examine what happens to the brackets, digits and alignment as the shopper moves from the product page to checkout.

Treat display text and machine values separately. A shopper can read Persian digits while an integration receives the original numeric representation. Likewise, two spellings may match in search without rewriting the title that appears on the receipt.

What it does

  • Set the document's language and reading direction.
  • Check logical spacing when switching between Persian and English.
  • Isolate an inline product code so its punctuation reads correctly.
  • Normalise selected letters and digits in search text.
  • Compare the currency unit shown to shoppers with the stored amount and unit.

Install and use

  1. Give the agent permitted storefront templates and invented product data. The optional local text normaliser needs Node 20 or later.
  2. Ask, “Review this product name, its Latin stock code and the checkout total. Show me what happens in both languages.”
  3. Test the edited interface and compare its underlying identifiers, money values and timestamps with the original records before accepting a change.

FAQ

Source

Made by Naser Sadrmanesh. Original skill: rtl-and-persian-localization, published on GitHub under the MIT licence.

View on GitHub