machinewrapped/llm-subtrans

Open Source project using LLMs to translate subtitles (SRT, SSA/ASS, VTT)

What it solves

LLM-Subtrans is an open-source subtitle translator that allows users to translate subtitle files between any language pair supported by a chosen Large Language Model (LLM). It simplifies the process of using AI for translation by providing a structured way to handle subtitle formats and API integrations.

How it works

The application divides subtitle files into batches and sends them to a translation service via API. It supports a pluggable system for subtitle formats, currently including .srt, .ssa/.ass, and .vtt files. Users can choose from various providers, including OpenRouter, Google Gemini, OpenAI, DeepSeek, Anthropic, Mistral, and Amazon Bedrock, or connect to a custom server with an OpenAI-compatible API (such as LM Studio).

Who it’s for

Content creators, translators, and viewers who need to translate subtitles for videos, as well as researchers interested in using different LLMs for translation tasks.

Highlights

  • Multi-Provider Support: Integrates with a wide array of AI providers including OpenRouter, Gemini, OpenAI, DeepSeek, Claude, and Mistral.
  • Flexible Interface: Offers both a graphical user interface (GUI) for ease of use and a command-line interface (CLI) for advanced control.
  • Project Persistence: Uses .subtrans project files to track progress and allow users to resume interrupted translations.
  • Contextual Translation: Supports providing movie names, descriptions, and terminology maps to ensure consistent and accurate translations across a batch.
  • Format Conversion: Ability to read one subtitle format and write another (e.g., converting .ass to .srt) during the translation process.

Related

  • Project
  • Project
  • Project
  • Project
  • Project