Whisply

What Whisply is

Whisply is free software for transcribing, translating, annotating and creating subtitles from audio and video with Whisper models. It provides both a command-line interface and a browser app that runs locally. Batch processing accepts individual files, folders, URLs and lists of inputs. Processing stays on the user’s hardware, with implementations selected for CPU, Nvidia GPU or Apple Silicon MLX. Exports include JSON, RTTM, TXT, HTML, SRT, VTT and WebVTT. Word-level speaker annotation uses whisperX and pyannote, requiring a Hugging Face access token and acceptance of the pyannote model conditions. The project describes Whisply as a research tool for work such as qualitative interviews and group discussions. It is MIT licensed, permitting use, copying, modification, publication and distribution under the stated conditions. Setup requires FFmpeg and Python 3.10 through a version below 3.14. The README notes that overlapping speakers can reduce diarization accuracy and that built-in word alignment supports ten languages.