OpenReader

Migrates user data to the server, improves media caching and loading states, and fixes several TTS and storage playback issues.

Introduces speech-sdk integration for TTS providers, transitions to a modular monorepo architecture, improves worker monitoring, and fixes PDF highlight rendering and session management.

Introduces web page article imports, a custom document creator, SSRF prevention, and various UI and performance improvements.

Added Supertonic local speech integration, WAV/OGG audio support, and improved EPUB playback, voice discovery, and word highlighting.

Introduces multilingual reader and speech support, a background tasks engine with admin UI, and improved user data management and concurrency locking.

This release fixes Docker document previews, improves upload reliability for .docx files, enhances drag-and-drop previews, and optimizes PDF parsing and reader behavior.

Improved runtime packaging, added recovery for legacy PDF parse records, and updated the filesystem cleanup process.

Improved PDF parsing resilience and enhanced reader UI with better progress feedback and job recovery.

Fixed a Docker container startup crash by switching the base image to Debian slim and optimized model prewarming defaults.

Fixed PDF parse state persistence for worker-backed systems and removed legacy runtime flags and related UI elements.

Fixed a major TTS control loading bug in the player UI and updated the compute worker deployment documentation.

Version 4 introduces layout-aware PDF parsing, a worker-backed compute architecture, a redesigned library UI, and a required authentication model.

Introduces a new Admin panel for managing TTS providers and site features, alongside improved multi-page TTS preloading, caching, and per-document tuning settings.

Introduces the Replicate TTS provider, improves TTS generation and playback controls, updates document library naming, and enhances deployment stability via SeaweedFS pinning.

Introduced persistent custom themes with live color editing and a new ColorPicker component, along with a configurable timeout for TTS upstream requests.

Refactored TTS provider handling and audiobook generation flow, improved configuration persistence, and enhanced system stability with bug fixes.

Improved TTS pipeline reliability, refactored audiobook generation and export, and introduced new UI themes and settings tabs.

Added official KittenTTS-FastAPI support as an OpenAI-compatible TTS option and improved documentation deployment workflows.

OpenReader 2.0 introduces user authentication, TTS rate limiting, S3-compatible storage, a new database layer, and a full rebrand from OpenReader-WebUI.

This release improves PDF highlighting reliability and revamps text normalization and TTS block splitting for better content handling.

Introduces external library imports, enhanced audiobook export settings, a document selection modal, and improved API security and Docker stability.

Added file previews to the document grid list and updated packages to address React and Next.js security vulnerabilities.

This release includes bug fixes for the audiobook export modal.

Introduces word-by-word highlighting with whisper.cpp, bundles the Whisper binary in Docker, and improves EPUB navigation and API structure.

Added text highlighting for the current sentence during EPUB playback with a new configuration toggle.

Version 1 introduces smart sentence continuation, a modernized PDF highlighting pipeline, audiobook exports, Dexie-backed local storage, and various backend and accessibility improvements.

Introduces multi-provider TTS support, migrates to pnpm, and improves Docker configurations.

This update enables internal link clicking, multiple file uploads, automatic page selection, and adds loading indicators for document selection and progress tracking.

Added PDF location saving, player and voice speed controls, and improved the SettingsModal layout.

Added editable page numbers to the PDF navigator view and refactored NLP into a shared function.

Version v0.3.0 adds support for .txt files and GitHub-style Markdown rendering via remark-gfm.

Added support for Orpheus-FastAPI and improved PDF text extraction space detection and TTS page-change pausing.

Added support for gpt-4o-mini-tts, custom model names, and .docx files, along with a redesigned settings modal.

Introduced TTS API retry logic, Playwright E2E testing, and updated PDF and audio files to use array buffer storage.

Moves the progress bar to a persistent popup, switches to GitHub container registry, and fixes an issue with longer audiobooks.

Added ePub and PDF export to audiobooks, custom PDF text extraction margins, and a high-precision voice speed slider.

Updated Next.js server environment variable handling, fixed background pausing, and reintroduced the AbortController API for request management.

TTS requests are now handled by the server instead of the client, alongside major bug fixes for iOS playback and API preloading.

TTS requests are now handled by the next server instead of the client, accompanied by various bug fixes and playback improvements.

Introduces an experimental themed ePub viewer, improves PDF stability via debouncing, and refactors ePub navigation.

Introduces new themes, renames the aqua theme to ocean, and includes various UI and bug fixes.

Introduced folders in documents list with drag-and-drop creation, file sorting, and updates to the settings modal.

Added new theme options and functionality to delete documents from local and server via a new endpoint.

Added server-side document storage for settings and improved NLP sentence splitting for ePub files.

Improves EPUB location saving, PDF state handling, sentence processing, and introduces automated Docker image updates via GitHub Actions.

Added EPUB support via react-reader and implemented bug fixes for faster sentence playback.

This update includes bug fixes, major refactors, and an update to the demo server's Kokoro-FastAPI version.

Introduced new viewing modes (single page, dual page, continuous scroll), UI improvements, and various bug fixes.

The app is switching from a long scrolling PDF view to a single page view with automatic page turning during TTS.

Introduced React window for PDF page virtualization to significantly improve performance and highlight auto-scroll accuracy.

Initial release featuring PDF upload, OpenAI API integration for LLM and TTS, theme support, and IndexedDB storage.