Translate Books with LLMs is an open-source tool that translates full-length books and long documents with large language models while preserving the original formatting. The GitHub project has collected around 2,540 stars and is actively updated in 2026. It handles EPUB, DOCX, PDF, and subtitle files, resumes interrupted jobs where they left off, and imposes no file size limits, which makes it practical for entire novels and technical manuals.
The tool chunks a document intelligently, sends each part to your chosen model provider, and reassembles a translated file that keeps headings, structure, and styling intact. It supports local models through Ollama and LM Studio as well as cloud APIs from OpenAI-compatible services, Gemini, Mistral, DeepSeek, Poe, and OpenRouter, so you can trade cost, privacy, and quality per project. A simple graphical interface makes it usable by non-programmers, and its resume capability means a crashed or paused job never wastes completed translation work.
The project pays attention to the practical details of book-length translation: a glossary keeps names and terminology consistent across chapters, chunk boundaries respect document structure so sentences are not split mid-thought, and progress is saved continuously rather than only at the end. Support for subtitle files also makes it useful for translating films and lecture series. Its active maintenance and responsive issue tracker have made it a go-to recommendation in ebook and language-learning communities.
The tool is free and open source under the AGPL-3.0 license. With local models it costs nothing at all; cloud providers charge their normal API rates.
Readers, translators, and small publishers who want to translate whole books and documents affordably while keeping the layout, and privacy-conscious users who prefer local models.