@mapleshadow

faster-whisper-zh provides local speech-to-text transcription with gpu acceleration, word-level timestamps, and multiple model sizes. covers chinese and multilingual audio via command-line interface with json and text output formats.

markitdown-zh wraps microsoft's markitdown library to convert pdf, docx, pptx, xlsx, csv, json, html, epub and other formats to markdown. supports batch processing and image extraction via python scripts or direct api calls.