0.6.0

feature

• Added a transcribe-then-summarize workflow that creates reusable transcript files before summarization.
• Added long-audio transcription with time-based chunking, parallel chunk processing, temp file updates, and chunk retry handling.
• Added category classification and category prompts for transcript summarization.
• Added transcription-only mode and consolidated transcription mode settings.
• Added transcript file links in generated output and removed embedded transcript links.
• Added cached upload reuse, upload metadata logging, and the gemini-3.5-flash model option.
• Removed the deprecated plain-text API key setting in favor of secure API key storage.

fix

• Fixed cancellation and retry cleanup across transcription, chunk retry, classification, and summarization flows.
• Fixed quota error handling and temporary transcription file cleanup.
• Fixed Obsidian reload/resource cleanup issues that could trigger offref errors.
• Stabilized settings loading and propagated Obsidian file write failures to callers.