DEV Community

ox3adie1
ox3adie1

Posted on

How I Built a Browser-Based UDF to PDF Converter for Turkish Legal Documents

#pdf #opensource #javascript #webdev

The Problem: A Proprietary Format Locked in a Government System

Turkey's judiciary runs on UYAP (Ulusal Yargi Agi Platformu) — the National Judiciary Informatics System. Every court document, petition, and ruling passes through it. And every one of those documents is stored in a proprietary format called UDF (UYAP Document Format).

Here's the catch: you can't open UDF files outside UYAP. There's no standard viewer, no export button, no simple way to share these documents as PDFs. Lawyers receive .udf files and have to jump through hoops to convert them.

I decided to fix that.

Meet UDF to PDF Converter

convertudftopdf.com is a free, open-source tool that converts .udf files to .pdf format — entirely in the browser.

No server uploads. No registration. No file size limits. Just drag, drop, and download.

What's Inside a UDF File?

This was the first mystery to solve. After some reverse engineering, I found that UDF files are actually ZIP archives containing:

document.udf (ZIP)
├── content.xml      ← The actual document content (RTF inside XML)
├── metadata.xml     ← Document metadata
└── [attachments]    ← Optional embedded files
Enter fullscreen mode Exit fullscreen mode

The content.xml file wraps RTF-formatted text inside XML nodes. So the conversion pipeline looks like:

.udf → ZIP extraction → XML parsing → RTF extraction → PDF generation
Enter fullscreen mode Exit fullscreen mode

The Technical Challenges

1. Client-Side ZIP Extraction

Using JSZip, extracting the ZIP is straightforward:

const zip = await JSZip.loadAsync(file);
const contentXml = await zip.file('content.xml').async('string');
Enter fullscreen mode Exit fullscreen mode

2. RTF Parsing

RTF is a surprisingly complex format. I needed to handle:

  • Character encoding — Turkish characters have special Unicode handling
  • Formatting — Bold, italic, underline, font sizes
  • Tables — Court documents are full of tables
  • Page breaks — Multi-page documents need proper pagination

I built a custom RTF parser that walks through control words and groups, extracting text with formatting metadata.

3. Turkish Font Support

This was the trickiest part. Default PDF fonts don't support Turkish characters properly. The dotless i and capital I with dot are particularly problematic.

The solution: embedding Noto Serif font directly into the generated PDF using jsPDF's font embedding API.

doc.addFont('NotoSerif-Regular.ttf', 'NotoSerif', 'normal');
doc.setFont('NotoSerif');
Enter fullscreen mode Exit fullscreen mode

4. Privacy by Design

Legal documents are sensitive. Lawyers can't upload court files to random servers. That's why everything runs client-side:

  • File reading: FileReader API
  • ZIP extraction: JSZip (in-browser)
  • PDF generation: jsPDF (in-browser)
  • Zero network requests during conversion

Architecture

The entire app is static HTML + CSS + JavaScript, hosted on GitHub Pages:

├── index.html          ← Main converter page
├── converter.js        ← Core conversion logic
├── style.css           ← UI styles
├── fonts/              ← Embedded Noto Serif font files
├── en/                 ← English version
├── udf-to-pdf.html     ← SEO landing page
└── karsilastirma.html   ← Tool comparison page
Enter fullscreen mode Exit fullscreen mode

No framework. No build step. No backend. It just works.

Results

Since launching, the tool has been used by lawyers across Turkey to convert thousands of documents. The feedback has been overwhelmingly positive — turns out, solving a real pain point with a simple tool goes a long way.

Try It Out

If you work in legal tech or deal with proprietary document formats, I'd love to hear about your experience. And if you find the tool useful, a star on GitHub would be appreciated!

Have you dealt with proprietary document formats in your projects? How did you approach the conversion? Let me know in the comments!

Top comments (0)