ReadAny: a local-first AI e-book reader with RAG-powered chat and semantic search

What it solves

ReadAny is a local-first AI e-book reader designed to prevent the loss of information during reading. It addresses the problem of scattered notes, keyword-only search limitations, and the difficulty of extracting specific answers from long texts by integrating AI directly into the reading workflow.

How it works

The application uses a Retrieval-Augmented Generation (RAG) architecture. It creates a local vector store and uses local embeddings (via Transformers.js) to enable semantic search and AI chat. Users can connect their own LLM providers (such as OpenAI, Claude, Gemini, or local models via Ollama) to ask questions about their books, with answers grounded in the reading context. It also supports WebDAV for cross-device synchronization of libraries and notes.

Who it’s for

Readers who want to build a private, AI-enhanced knowledge base from their e-books across desktop and mobile platforms, and those who prioritize data privacy by keeping their vector stores and embeddings local.

Highlights

• AI-Native Reading: Includes semantic search (hybrid vector + BM25), intelligent chat grounded in current reading position, and a built-in "skills system" for summarization and concept explanation.
• Local-First Privacy: Uses local embeddings and a local vector store to keep data offline-capable.
• Comprehensive Format Support: Compatible with EPUB, PDF, MOBI, AZW, AZW3, FB2, FBZ, CBZ, and TXT.
• Knowledge Management: Features 5-color highlights, Markdown notes, and exports to Obsidian, Notion, HTML, and JSON.
• Cross-Platform: Available for macOS, Windows, Linux, iOS, and Android.
• Accessibility: Integrated Text-to-Speech (TTS) with over 100 voices via Edge TTS, Browser TTS, and DashScope.