Modernizing BookReader: Neural TTS integration and Native Browser PDF Support #1476
Description
Is your feature request related to a problem? Please describe. The current BookReader UI and Text-to-Speech (TTS) engine feel outdated compared to 2025 web standards.
The TTS voice is robotic/mechanical and lacks natural prosody.
The page-flipping animation/scrolling is not as fluid as modern browser-native PDF rendering.
Users frequently encounter "0 views" on items, which seems to be a metadata sync lag/bug.
Describe the solution you'd like I suggest a modular update to the BookReader:
Neural TTS Integration: Instead of the legacy speech engine, integrate the Web Speech API or lightweight WASM-based neural modules (e.g., Piper or similar). This allows for lifelike voices and automatic language switching without heavy server-side processing.
Native Rendering Option: Allow the reader to leverage native browser PDF rendering capabilities for smoother scrolling and better performance on modern devices.
Stat Visibility: Implement a "hide if zero/error" logic for view counters if the statistics server is unreachable, to avoid misleading users.
Describe alternatives you've considered Allowing users to bypass the BookReader entirely in favor of the browser's built-in viewer, which already includes accessibility features and high-quality "Read Aloud" functions (like in Edge or Chrome).
Additional context Improving these "front-facing" features will significantly enhance the user experience for researchers and the visually impaired, keeping the Archive competitive with modern digital library standards.