Batch Subtitles with whisper.cpp

whisper.cpp_batch_subtitle is a PowerShell script for turning a directory of videos into subtitle files with FFmpeg and a local whisper.cpp build. It keeps media and transcription on the user's machine and applies the same command-line settings across a batch.

The setup has three external pieces: install FFmpeg and place it on PATH, build whisper.cpp for the available CPU or GPU, and download an appropriate Whisper model. Then edit $whisperPath, $modelPath, and any whisper.cpp CLI arguments in whisper.cpp_transcribe.ps1 before running the script and selecting the video directory.

Model size and hardware backend determine speed, memory use, and accuracy. Language selection, timestamps, and output format come from the whisper.cpp CLI rather than this wrapper, so test one representative file before processing a large folder.

The repository notes that this workflow has been superseded by the Python transcription command in ExplicitUtil, which uses asynchronous processing and is the better starting point for new work. The PowerShell version remains useful as a compact Windows-focused reference.

Visit the whisper.cpp batch subtitle repository for the script and current setup notes.