Any Subtitle

Generate local live or accurate captions with automatic multilingual speech recognition.

Any Subtitle displays locally generated captions in the spoken language over the current video player. - Live captions update every few seconds from user-selected tab audio. - Accurate captions use existing page tracks when available or transcribe accessible media with a larger Whisper model. - Multilingual recognition automatically detects and transcribes spoken languages supported by Whisper. - FFmpeg, yt-dlp and CUDA whisper.cpp run through a local Windows companion; there is no cloud speech recognition. - Accurate tracks are cached locally for seven days. - Captions follow play, pause, seek, playback speed and fullscreen changes. First use requires the free Any Subtitle Local Core for Windows. The setup page provides the download and checks its status. The complete installation needs about 5 GB of free space and an NVIDIA GTX/RTX GPU. Any Subtitle does not provide video download or export functionality. Use it only with content you own or are authorized to access. DRM media, canvas-only players, browser internal pages and Picture-in-Picture windows are not guaranteed.