Add local LLM post-processing workflow

[agata]

Summary

• integrate a local LLM post-processing engine that runs Whisper transcripts through configurable prompts, dictionary hints, and diff logging
• surface post-processing controls in the settings UI with presets, custom prompt management, history browsing, and new status handling
• document the workflow with dedicated English/Japanese guides and extend i18n strings to cover the new LLM features

Testing

• cargo fmt --all
• cargo build