Overview
Pitaka is a desktop app for reading and researching an EPUB and PDF library. Import your books, search them all at once by word or by meaning, collect passages into bookmark folders, and highlight and write notes as you read. Everything runs locally — your books never leave your machine.
There’s a browser demo with a built-in public-domain book, the Therīgāthā (Verses of the Senior Nuns), and installers for Windows and Linux on the releases page.
Features
- Search across the whole library — ranked full-text search with highlighted matches. Case and diacritics are ignored, so “samsara” finds “saṃsāra”. Supports quoted phrases,
prefix*, andAND/OR/NOT. - Find chapters by meaning — describe what you’re looking for (“the peace of letting go”) and a small embedding model, running entirely on your own computer, finds the chapters that match.
- Reader — opens search hits in context, with a chapter list, reading progress, and resume-where-you-left-off.
- Highlights and notes — four highlight colours, with notes on a highlight or a whole paragraph shown in the margin.
- Bookmark folders — collect passages into named folders, one per topic or project.
- Display settings — text size, line spacing, column width, and light, sepia, or dark themes.
Technology stack
- Core: a Rust crate (
ebook_research_core) handling EPUB and PDF parsing, the SQLite database, and search, with no dependency on the UI shell - Search: SQLite FTS5 for full-text search; candle running
bge-small-en-v1.5locally for semantic chapter search - App shell: Tauri v2 with a thin command layer over the core crate
- Frontend: React and TypeScript, built with Vite
Pitaka is open source under the MIT or Apache 2.0 licence.