Transcribing fifty files one process at a time spends the majority of its runtime loading weights, and a persistent server removes that entirely. Pairing it with a batch mode that advances itself means a folder can be left to run. The setup and hardware checks earn their place for the same reason the batch does: whisper.cpp on Windows with Vulkan has several ways to be quietly not accelerated, and a transcription that is ten times slower than it should be looks identical to one that is working. Everything runs locally, so audio never leaves the machine — which for recordings of meetings or interviews is usually the whole reason to pick this over a hosted service.
A local Whisper transcription client for Windows built around whisper.cpp: single and batch transcription, subtitle generation, model management, and hardware verification.
- check_config verifies whisper-cli.exe, the model and FFmpeg are all present, and its own text says to run it first — the setup check before anything is attempted.
- check_system detects the GPU and verifies Vulkan acceleration is available, so a slow transcription is explicable rather than mysterious.
- transcribe_audio runs a single file as a background job with check_progress following it, and generate_subtitles produces subtitle files with the language settable.
- Two batch modes: an interactive one that goes file by file, and start_batch which runs sequentially through everything untranscribed with check_batch_progress advancing it automatically.
- Models managed in place — list what is installed with file sizes, download one from Hugging Face into the models directory, and switch the active model without restarting.
- analyze_media runs FFprobe on the files before transcribing, which is how you find out a file is not what it claims before spending time on it.
- whisper_server starts, stops and checks a persistent model server, so the model is not reloaded per file.
Windows with whisper.cpp, FFmpeg and a Whisper model — WHISPER_CLI_PATH and WHISPER_MODEL point at them, and check_config verifies all three.
One command — npm install -g whisper-windows-mcp