blog-translate

Blog Translate converts existing blog posts into one or more target languages with SEO-optimized localization rather than generic machine translation. It solves the problem that naive translation breaks page structure and loses search performance: it preserves markdown, MDX, and HTML structure, frontmatter, schema JSON-LD, and image and chart embeds, while producing native-quality, publication-ready copy. It localizes keywords, meta tags, numbers, dates, currencies, and quote styles per locale, and flags machine-translation artifacts for review. It is meant to run before the companion blog-localize skill: translate handles language conversion, localize handles deeper cultural adaptation.

The workflow parses input safely (resolving paths against the project root, rejecting symlinks, traversal outside the root, files over 10 MB, and binary files), auto-detects source language, and normalizes target codes to ISO-standard hreflang tags with validation and suggestions. It then extracts the translatable surface (headings, body, alt text, figcaptions, SVG text, FAQ, CTAs) while preserving URLs, code fences, internal-link markers, and names of people, brands, and organizations. Keyword localization decides per language whether to keep an established loanword or swap to a locally-searched equivalent. Reference files provide per-locale number/date/currency/quote rules, SVG length-delta guidance, a quality checklist, and cultural adaptation profiles for DACH, Francophone, Hispanic, Japanese and other markets.

It is aimed at content and SEO teams publishing multilingual blogs who need structurally faithful, search-optimized translations. Typical uses include translating a post to German, French, Spanish, or Japanese while keeping schema and chart markup intact.