A single Win10 executable. Press Alt+X anywhere, drag a rectangle, get the recognized text appended to today's YYYY-MM-DD-ocr.md file — fully offline, CPU-only, tray-resident, with optional boot-time ...
The tool uses inaSpeechSegmenter (a CNN-based speech/music/noise classifier) to segment each file, finds the end of the last music segment, and cuts everything after it using ffmpeg. Music in the ...