A fast AI-assisted PDF reader for papers.
- Frontend: SvelteKit, Svelte 5, TypeScript, Vite
- Desktop shell: Tauri v2
- Native core: Rust 1.85.0 baseline
- PDF engine: MuPDF through
mupdf-rs - Translation: OpenAI-compatible streaming chat completions
- Open local PDFs with a native file dialog.
- Render pages through MuPDF via the
jrpage://Tauri custom protocol. - Use continuous virtualized scrolling with render prefetching and an LRU page-image cache.
- Overlay a transparent text layer for mouse selection.
- Translate selected text in a resizable right or bottom panel.
- Configure base URL, model, API key, language pair, and prompt template.
- Store settings in the OS app config directory and API keys in the OS keychain.
npm install
npm run tauri devnpm run check
npm test
npm run build
cd src-tauri
cargo test
cargo clippy --all-targets -- -D warningsThis project is licensed as AGPL-3.0-or-later to match the open-source MuPDF dependency model.