What's New

This page summarizes the most recent user-facing updates now reflected in the website docs. The changes below cover the latest feature work in v0.4.12 through v0.4.14.

v0.4.14: document tags and library workflows

Reader now includes a complete document tag system.

New capabilities

• real persisted tags instead of display-only inferred labels
• tag filters in the Library sidebar
• Any / All matching for multi-tag filtering
• current-document tagging in the Tags panel
• AI-generated tag suggestions with review actions:
  • Accept
  • Create Temp
  • Map
  • Reject
• Related Documents discovery based on shared tags

New management workflows

• Batch Tags for bulk updates across many documents
• Tag Library for Rename, Merge, Add Alias, Promote, and Cleanup Unused

Why it matters

This turns Reader's library organization into a reusable taxonomy instead of a one-off UI grouping system.

v0.4.13: reading comprehension and glossary workflows

Reader's selection actions were reworked into a more focused understanding flow.

New Understand panel

The right-side tool panel now includes:

• Explain Simply
• With Context
• Term
• Takeaway

Each action is optimized for a different reading problem:

• simplify hard language
• include nearby context
• analyze important terminology
• extract a study note

New glossary system

Reader now stores local glossary entries and lets you:

• save a Preferred Rendering
• store Concept Tags
• browse entries in the Glossary panel
• reuse saved terminology choices in later Term analysis

Why it matters

Reader is no longer limited to one-off explanations. It can now accumulate terminology decisions and reuse them while you keep reading.

v0.4.12: text parse and multimedia parse modes

Reader added more flexible Markdown reading modes.

New reading view controls

The Reading View menu now groups:

• Text Parse
• Multimedia Parse
• Source Only
• Translation Only
• Source + Translation

What changed

• Text Parse keeps Markdown reading focused and text-first
• Multimedia Parse improves handling for Markdown content with remote article assets, linked media, and inline image placeholders
• bilingual mode controls are now grouped with the rest of the reading view options

Why it matters

Markdown documents can now be read in a way that better matches the source material instead of forcing a single rendering strategy.

Also included around the same release window

• stronger AI profile management with separate provider, model, and agent configuration
• tag-aware library cards and refreshed document metadata display
• continuing support for translation, semantic search, notes, TTS, and MCP tooling

Next

• User Manual
• Installation
• MCP Tools Reference