uzbek transliterate

Standard letter-by-letter transliteration fails on Uzbek orthography, producing errors like "администратсия" instead of "администрация" or losing soft signs in month names. These mistakes are immediately obvious to native speakers and undermine the quality of localized software and reports. This skill provides a specialized transliteration engine that follows official grammar rules and handles complex loanword exceptions.

What it does

• Loanword handling correctly maps Russian and international terms using a 1,000+ word dictionary to preserve characters like ц, ь, and э.
• Bi-directional conversion transforms text between Latin (uz-Latn) and Cyrillic (uz-Cyrl) scripts accurately.
• Smart apostrophe detection recognizes multiple Unicode variants of apostrophes for o‘ and g‘, ensuring consistent output regardless of source input.
• Case preservation maintains ALL-CAPS formatting for names and acronyms while applying script-specific casing rules.
• Suffix logic identifies Uzbek suffixes to apply loanword stems correctly while dropping soft signs before inflection.

How it works

1. Command line uses node scripts/translit.mjs --to cyrl "text" for immediate single-string conversion.
2. Bulk processing pipes multiple lines through stdin to convert large lists of values or labels while maintaining order.
3. File mode applies script changes to entire plain-text files using the --file and --out flags.
4. Programmatic import uses the toCyrillic and toLatin functions directly in JavaScript or TypeScript projects.

Frameworks & tools

The skill is built with Node.js (version 18+) and has zero external dependencies. It includes a JavaScript port of the LotinKrill logic and integrates easily with localized JSON, .resx, or CSV workflows.

Why this beats prompting it yourself

Generic LLM prompts often hallucinate apostrophe usage or miss subtle orthography rules like the ц/s distinction based on vowel position. This tool uses a deterministic dictionary-based approach to ensure "oktabr" becomes "октябрь" every time without probabilistic errors.

Use cases

• App localization to generate Cyrillic .json or .resx strings from Latin source files.
• Data normalization for converting lists of names and addresses in Excel or CSV exports.
• Content migration to move blog posts or documentation between script standards.

Known limitations

Source typos are preserved. Words missing from the internal loanword dictionary are converted letter-by-letter. The characters 'c' and 'w' are ignored as they are not standard Uzbek letters.