PDFRead

A Tauri desktop PDF bilingual reader with sentence-level alignment and translation.

Why PDFRead

Reading technical PDFs in another language is slow. PDFRead keeps the original PDF on the left and sentence-level translations on the right, so you can scan and compare without losing your place.

Highlights

• Two-column layout: PDF on the left, translations and controls on the right.
• Sentence-level translation with aligned highlights on the PDF.
• Read-only Slate rendering for clean, consistent text.
• Virtualized translation list for long documents.
• Local JSON cache for translations (no repeated costs).

How It Works

• PDF rendering: pdf.js (pdfjs-dist/legacy/build/pdf.mjs) with text layer overlays.
• Sentence extraction: single-column heuristic in src/lib/textExtraction.ts.
• Translation pipeline: frontend -> Tauri command -> OpenRouter -> cache.
• Cache keys: docId + sid + source text hash + model + target language.

Install (Homebrew)

brew update && brew tap everettjf/tap && brew install --cask pdfread

Develop

bun install
bun run tauri dev

Build

bun run build

Usage

1. Open a PDF.
2. Select target language and model.
3. Hover or click a sentence to highlight corresponding regions in the PDF.
4. Translations are cached locally for fast repeat access.

Settings

• Theme: system / light / dark
• Translation mode:
  • Window: current page ± radius
  • Chunk: translate by page blocks
• Debounce: 400ms; concurrency: 1; stale responses ignored

Project Structure

• src/components/PdfViewer.tsx: PDF list and navigation
• src/components/PdfPage.tsx: PDF page rendering and highlights
• src/components/TranslationPane.tsx: Translation UI (Slate read-only)
• src/lib/textExtraction.ts: Sentence extraction
• src-tauri/src/lib.rs: OpenRouter translation command

Recommended IDE Setup

• VS Code + Tauri extension + rust-analyzer

Star History