guaguastandup/zotero-pdf2zh

What it solves

Zotero PDF2zh is a plugin for Zotero that allows users to translate PDF documents directly within their research library. It solves the problem of having to manually export PDFs to external translation tools, preserving the original layout of the academic papers while providing translated text.

How it works

The project consists of two main components: a Zotero plugin (XPI file) and a Python-based backend server. The plugin communicates with the local Python server, which leverages translation engines like pdf2zh or pdf2zh_next to process the documents. Users can configure various LLM API providers (such as DeepSeek, SiliconFlow, or Google/Bing) to handle the actual translation. The system supports multiple output formats, including bilingual side-by-side layouts and cropped versions optimized for mobile reading.

Who it’s for

Academic researchers, students, and anyone using Zotero to manage scientific literature who needs to translate foreign-language PDFs into their native language (specifically Chinese) while maintaining the document's structure.

Highlights

• Multiple Translation Engines: Supports both the original pdf2zh and the more advanced pdf2zh_next for improved table translation and OCR compatibility.
• Flexible LLM Integration: Compatible with a wide range of translation services, including free options (SiliconFlow) and high-quality paid APIs (DeepSeek, Aliyun DashScope).
• Layout Preservation: Offers various output modes such as dual-language (Left & Right), Top & Bottom, and a specialized "crop-compare" mode for double-column papers.
• Integrated Progress Tracking: Includes a web-based interface to monitor translation progress and view translation history.
• Batch Processing: Allows users to select multiple items in Zotero and translate them in bulk.