rockbenben/subtitle-translator

What it solves

Subtitle Translator addresses the common issues of using general-purpose translators for subtitle files: the risk of AI models rewriting or corrupting timecodes and the inefficiency of translating files one by one. It ensures the timeline remains untouched by separating dialogue from timing data locally before sending text to the translation engine.

How it works

The tool is a browser-based application that processes subtitle files (.srt, .ass, .vtt, .lrc) locally. It strips the timing and cue numbers from the files, sending only the dialogue text to a configured translation API. After translation, it re-inserts the timing data to reconstruct the subtitle file.

It supports two main translation paths: traditional Machine Translation (MT) APIs (like DeepL and Google Translate) and LLM providers (like OpenAI, Claude, and DeepSeek). LLM modes offer "context-aware translation," where surrounding lines are sent as context to improve dialogue coherence and character voice consistency.

Who it’s for

Content creators, anime fansubbers, and anyone needing to translate entire seasons of shows or movies into multiple languages while preserving the exact timing and formatting of the subtitle files.

Highlights

• Batch Processing: Translate hundreds of files at once with parallel processing for high speed.
• Privacy-First: All parsing, API key storage, and requests happen client-side in the browser.
• Bilingual Output: Ability to generate subtitles with both original and translated text, including specialized ASS styles.
• Format Conversion: Built-in conversion between SRT, VTT, and ASS formats during the translation process.
• Context-Awareness: LLM-driven translation that uses surrounding lines to maintain narrative consistency.
• Extensive API Support: Compatible with 7 traditional MT APIs and over 17 LLM providers.