As someone who reads dozens of ArXiv papers, technical specs, and digital books daily, reading dense multi-column pages in my second language was draining my brain. The Problem: Existing translation tools drove me crazy. Google and DeepL break multi-column layouts, overlapping text and making pages unreadable. Dropping full PDFs into ChatGPT/Claude burns tokens insanely fast, merges columns randomly, and turns math formulas into broken gibberish. How I Solved It: I built PdfTranslator for iOS—a lightweight app designed to translate PDFs without ruining their structure. 1. 1:1 Layout Preservation: ArXiv papers, scanned books, and manuals keep their original visual structure. Multi-column text, tables, and inline math formulas stay exactly where they belong. 2. Smart Text Auto-Fitting: Dynamically adjusts font sizes and line heights so translated text fits the page gracefully without overflowing or overlapping neighboring elements. 3. Unlimited On-Device Mode: Uses local OCR and native iOS translation for 100% offline privacy, zero server costs, and zero page limits. 4. Smart LLM Mode: Passes structured text to AI for high-precision jargon translation without wasting tokens on layout overhead. I built this tool to solve my own daily headache, and it has completely changed my reading workflow.

PdfTranslator

Translate papers & books without losing 1:1 layout

Translate papers & books without losing 1:1 layout