A document translator that keeps the original design

Translating PDF files without messing up the formatting is notoriously difficult. It's a problem people have been trying to solve for a long time. There are some major options available, but they have clear limitations: DeepL and Google Translate: Their translation quality is excellent, but their native document translation features often completely break the formatting in complex PDFs. Microsoft Azure Translator: It's also a powerful engine, but you're stuck in their ecosystem, and handling the document formatting still takes a lot of effort. I built DocTranslating to solve this. First, it brings together the main APIs (DeepL, Google Translate, and Microsoft Translator) into a single platform, letting you switch between them seamlessly depending on which engine handles your specific language pair better. I also built my own custom workflow to translate PDFs. It uses Google's Gemini API to handle the heavy lifting of text translation, but the document processing and layout reconstruction happen entirely on our servers. The PDF processor is built on multiple open-source (FOSS) tools to meticulously extract text and put it back without messing up the original structure. short video demo: https://youtu.be/YeAG3Om9Xg8?si=6zr7FyzAN3oBMgtb