ReadAny is an open-source, local-first e-book reader designed to help users query their reading material through semantic search and AI-driven interaction. Built for desktop (macOS, Windows, Linux) and mobile (iOS, Android), it uses a RAG pipeline with hybrid retrieval—combining vector search and BM25—to allow users to find ideas by meaning rather than just exact keyword matching. The application prioritizes privacy and flexibility by running embeddings locally and allowing users to connect various model providers like Ollama for local execution or OpenAI, Claude, and Gemini via API.
- Supports more than ten book formats with note export available in five different formats.
- Includes features such as Text-to-Speech (TTS), reading statistics, a skills system, and WebDAV sync for multi-device use.
- Offers high flexibility by supporting various model providers including DeepSeek and custom-compatible endpoints.
- Supports more than ten book formats with note export available in five different formats.
- Includes features such as Text-to-Speech (TTS), reading statistics, a skills system, and WebDAV sync for multi-device use.
- Offers high flexibility by supporting various model providers including DeepSeek and custom-compatible endpoints.
Calibre levels up with an AI feature and other changes. This release introduces the 'Ask AI' tab in the e-book viewer, allowing users to highlight text and ask AI models questions. It also includes bug fixes and usability improvements.
