Built a local AI subtitle generator for converting audio and video files into transcripts and timed SRT subtitles. The app uses Python and CustomTkinter for a polished desktop interface, Whisper/OpenAI Whisper with MLX acceleration on Apple Silicon, and FFmpeg/FFprobe for media probing and audio preprocessing. It supports single-file and folder batch processing, configurable transcription models, denoise/normalization profiles, optional local LLM-based technical term refinement, progress tracking, stop-after-current-file control, diagnostic logging, settings persistence, and desktop packaging through PyInstaller with platform-specific build and publish scripts.