hydropix/TranslateBooksWithLLMs

What it solves

It simplifies the process of translating long-form documents—such as books, subtitles, and various file formats—while maintaining the original formatting and ensuring linguistic consistency across the entire work. It overcomes the common LLM limitation of context windows by handling documents of any length without losing context between segments.

How it works

The application uses an intelligent chunking system to break large files into manageable pieces for LLMs. It supports both local providers (via Ollama or OpenAI-compatible servers) and cloud providers (OpenAI, Gemini, DeepSeek, etc.). To ensure a consistent tone, it can extract writing style presets from sample books or automatically derive a glossary and style from the document itself. It also includes a checkpoint system to resume interrupted translations and supports parallel processing for cloud-based translations.

Who it’s for

It is designed for authors, translators, and readers who need to translate EPUB, SRT, DOCX, or TXT files using AI while preserving the structural integrity of the original document.

Highlights

• Format Preservation: Keeps EPUB styles, structure, and SRT timecodes exactly as they were in the original.
• Flexible LLM Integration: Works with a wide range of providers, including local (Ollama) and cloud (OpenRouter, Mistral, etc.).
• Consistency Tools: Features automatic glossary and style extraction to maintain a consistent register and rhythm.
• Unlimited Length: Handles documents from a single page to thousand-page novels via intelligent chunking.
• Resumable Progress: Automatically saves checkpoints so users can pick up where they left off.