Translate scientific PDF documents while preserving mathematical formulas, charts, layout, and formatting.
Full-text bilingual PDF translation preserving math formulas, tables, charts, and annotations.
PDFMathTranslate is a document translation tool designed for researchers, academics, and translators working with scientific papers. It translates PDF text line-by-line while keeping mathematical notation, charts, annotations, and original layouts intact. You can run it locally, deploy it via Docker, or connect it directly to translation APIs like DeepL, OpenAI, or Ollama.
The tool extracts document content and runs it through your choice of translation provider, including Google, DeepL, OpenAI, MiniMax, or local Ollama models. It uses an adaptive typesetting engine with experimental OCR support to reassemble bilingual or translated PDFs with matched layouts. It offers multiple interfaces including CLI, GUI, MCP server integration, and a Zotero plugin.
It suits researchers, technical translators, and students who need accurate translations of scientific papers without losing complex mathematical formatting. It is less suitable for casual readers who prefer basic browser extensions or those who cannot manage API key configurations or local Python/Docker setups.
Keeps mathematical equations, figures, tables, and document annotations in their original positions.
Connects to translation backends including Google Translate, DeepL, OpenAI, MiniMax, and local Ollama instances.
Offers a command-line tool, interactive GUI, Docker image, MCP integration, and Zotero workflow support.
Regroups scanned paragraphs and applies adaptive typesetting to match original document layouts.
The software is free under the AGPL-3.0 license. The only direct costs come from API usage fees charged by providers like OpenAI or DeepL, or the hardware required to host local LLMs. Free public web demos are available with limited computing resources.
874 open remote AI-training roles, updated November 15, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
These reviewed AI tools do the same job instantly — from our scored directory.
Yes, you can run it offline by pairing the local CLI or GUI with an offline LLM backend like Ollama.
It supports Google Translate, DeepL, OpenAI, MiniMax, Ollama, and other compatible translation services.
See Zekai first in Google