# ReadAny

> Local-first AI e-book reader with RAG chat, semantic search, highlights, text-to-speech and cross-device sync.

ReadAny is an open-source e-book reader for desktop and mobile built by the ReadAny Team. It lets readers chat with their books using their own AI provider, search by meaning, annotate passages and sync their library across devices. The current release listed on GitHub is v1.3.6.

## What It Is

ReadAny is a local-first reading app that combines an e-book reader with AI assistance and a personal knowledge base. It opens EPUB, PDF, MOBI, AZW, AZW3, FB2, FBZ and CBZ files, and imports TXT and UMD by converting them to EPUB. Books, annotations and reading data are stored locally, and the site says nothing is sent to external servers unless you explicitly sync.

## AI Reading Workflow

The AI chat uses your current reading context, including your position, selected text, chapter information and highlights, to answer questions, summarize, explain concepts or translate. Search combines vector retrieval with BM25, so ideas can be found even when the exact keyword is missing. Local embeddings and a local vector store keep this workable offline. Users connect OpenAI, Claude, Gemini, Ollama, DeepSeek or a custom OpenAI-compatible API with their own keys. A skills system provides built-in skills such as a summarizer, concept explainer and character tracker, and supports custom skills.

## Notes, Listening and Tracking

Highlights come in five colors, and Markdown notes can be exported to Markdown, HTML, JSON, Obsidian or Notion. Text-to-speech uses Edge TTS, Browser TTS or DashScope, with 100+ voices, speed control and background playback. Reading statistics include heatmaps, trend charts and streak tracking.

## Sync and Setup

Library, highlights and notes sync through WebDAV, S3 or LAN, with automatic background sync. Builds are offered for macOS, Windows, Linux, Android and iOS (via TestFlight), plus a Homebrew cask for macOS. The desktop app uses Tauri with React and TypeScript, and mobile uses Expo. Source builds require Node.js 18+, pnpm 9+ and Rust.

## Features
- AI chat grounded in current reading context
- Hybrid semantic search (vector + BM25)
- Local vector store and embeddings
- Multiple AI providers including custom OpenAI-compatible APIs
- Skills system with built-in and custom skills
- AI or DeepL translation in 19 languages
- Five-color highlights and Markdown notes
- Export to Markdown, HTML, JSON, Obsidian and Notion
- Text-to-speech with 100+ voices
- Reading statistics with heatmaps and streaks
- Sync via WebDAV, S3 and LAN
- Support for EPUB, PDF, MOBI, AZW, AZW3, FB2, FBZ, CBZ, TXT and UMD
- Light, dark and sepia themes

## Integrations
OpenAI, Anthropic Claude, Google Gemini, Ollama, DeepSeek, DeepL, Obsidian, Notion, WebDAV, S3, Edge TTS, DashScope

## Platforms
WINDOWS, MACOS, LINUX, ANDROID, IOS, API

## Pricing
Open Source

## Version
v1.3.6

## Links
- Website: https://codedogqby.github.io/ReadAny/
- Documentation: https://codedogqby.github.io/ReadAny/support/
- Repository: https://github.com/codedogQBY/ReadAny
- EveryDev.ai: https://www.everydev.ai/tools/readany