Back to changelog
v0.4.31

Conversation progress and local model improvements.

Added collapsible reasoning and tool progress, and improved local model startup, validation retention, and long-conversation handling for clearer waits and smoother continued use.

What's new

  • Added collapsible reasoning and tool-content previews with generated character counts, elapsed time, and execution stages for clearer long-running task progress.
  • Added a cancellable waiting state while local models start, preserving drafts and attachments. The app prioritizes the last manually selected model on startup.
  • Reasoning now follows the interface language setting for a more consistent Chinese and English experience.
  • Fixed lost StartLux validation state after model-directory changes, migration, or restoration, reducing unnecessary repeat tests.
  • Improved when long conversations are summarized and how processing states appear, keeping progress aligned with the corresponding messages and conversation history.
  • Extended the waiting window for local models processing long contexts and disabled automatic retries for timed-out requests to reduce repeated computation.