whisper-drop
Drag an audio or video file onto the window. Get a transcript. That's the whole app.
Under the hood it runs OpenAI's Whisper models locally via whisper.cpp. No server, no account, no upload: the only thing it ever fetches over the network is the model file, once. Built for the recordings you would never want sitting on someone else's server, like client calls and confidential interviews.
Copy the text out, or save it as .txt, .srt, or .vtt when what you actually needed was subtitles.