PDF to DOCX Translation Automation

The client provides translation services to medical plan providers like WellCare and similar companies. They receive hundreds of PDFs daily that need to be translated. Previously, the process was fully manual—converting each PDF into a DOCX file while preserving formatting, then translating it. This was extremely time-consuming, especially when dealing with complex layouts and tables.

Outcomes

The Problem

The client had to manually handle each PDF—convert it into a DOCX file, fix formatting issues, and then perform translation. The biggest challenge was preserving the original structure, especially for documents with tables and mixed layouts. This process was repetitive, slow, and prone to errors, making it difficult to scale as the volume of documents increased.

Challenges

• Converting PDFs to DOCX while preserving formatting and layout
• Handling complex table structures in certain document types
• Different PDF formats required different parsing approaches
• Maintaining consistency across all generated documents
• Reducing manual effort without breaking existing workflow

Solution

I built a custom desktop application using Tkinter to automate the entire workflow. The interface is simple—the user selects a folder containing PDFs and a template folder, then runs the process with a single click. The system uses a predefined DOCX template with placeholders, ensuring consistent formatting across all documents. PDFs are parsed using PyMuPDF, and relevant data such as names, plans, and other fields are extracted and inserted into the template. For more complex PDFs with structured tables, I implemented custom parsing logic tailored to each document type. For simpler documents, I used the OpenAI API to extract required fields from raw text and map them into the template. The application generates a complete output folder with clean, structured DOCX files, ready for translation. This significantly reduced manual work and made the process scalable.

Tech Used