LLM Translator Github

Purpose-built to translate technical content with an LLM while preserving software-specific terminology and code syntax. It accepts source_text and a required target_language, optionally auto-detecting the source. The system outputs translated_text plus metadata such as detected_source_language. It targets developers, technical writers, product teams, and localization engineers who localize GitHub READMEs, documentation, issue templates, and contribution guides.
Conceptually it performs detection, translation with domain constraints, and post-edit refinement. A glossary enforces consistent terminology, and code blocks, JSON, and YAML snippets are preserved; non-text elements are left intact. The tool supports batch translation and per-item metadata to fit into CI/CD or localization workflows.
Unique value lies in strict code preservation, glossary-driven consistency, and seamless integration with repository content, enabling faster localization cycles while minimizing terminology drift. Core features include language pair support, glossary import, optional back-translation checks, and export options for plaintext or markdown. Use cases cover translating READMEs, changelogs, CONTRIBUTING guides, and API docs, validating multilingual content before release, and preparing localized assets for distribution.