Whisper CLI
Whisper is a CLI for audio transcription, with model sizes for different hardware and output in the original spoken language.
- 112,049 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The CLI reads recordings, downloads models and writes results to disk.
Install
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/whisperDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The CLI reads recordings, downloads models and writes results to disk.
This is third-party code. Review the repository files before installing.
What it does
Whisper is a CLI for recognizing speech from audio files using OpenAI models on your computer. Users supply recordings, choose a model and receive a transcript. The spoken language can be set explicitly or detected by the model. A separate mode translates non-English speech into English. This is the reference Python implementation with a PyTorch dependency; the `whisper` command does not need an API account. Several recordings can be passed to one command. Users retain the source audio files.
Who it is for. Interview and teaching-material authors turning existing audio recordings into editable text.
Good fit when
- An editor transcribes several recordings in one terminal invocation.
- A teacher obtains a talk transcript in its original language.
- An author translates foreign speech into English with a multilingual model.
- A team matches model size to available memory before processing an archive.
Not a fit when
- The translator expects direct speech translation into Russian rather than English.
- A user chooses `turbo` for speech translation.
- The workflow needs a meeting recorder with a notes-search interface.
Example request
Transcribe this recording locally in its original language and flag passages that need listening again.Limitations
Audio processing requires a separate ffmpeg installation; the Python package does not replace it. Memory needs depend on the model; use its requirements table to choose a variant that fits your computer before processing a recording archive. The `turbo` model is not trained for translation and returns the source language even with `--task translate`. Recognition performance varies substantially between languages, so transcripts need comparison with the recording before publication.
How to disable. Stop transcription and remove `whisper` from jobs. Generated text and downloaded models remain on disk after the command exits.
FAQ
How do I install the CLI?
Run `pip install -U openai-whisper` and install ffmpeg separately for your operating system. The `whisper` command can then read audio; start with a short recording.
Can I process several files?
`whisper audio.flac audio.mp3 audio.wav --model turbo` accepts several recordings. Replace the filenames and inspect each transcript.
How do I specify the language?
The README example is `whisper japanese.wav --language Japanese`. Substitute your file and supported language instead of relying on detection.
How do I translate into English?
Use `whisper japanese.wav --model medium --language Japanese --task translate`. Select a multilingual model and substitute the source language.
Why does setup require Rust?
This can happen when tiktoken has no prebuilt wheel for your platform. Follow Setup’s Rust link; for `No module named 'setuptools_rust'`, the author gives `pip install setuptools-rust`.
Discussion
No questions or reviews yet.
Related
CLI toolswhisper.cpp is an audio CLI without a Python environment, compiled for a selected processor or hardware accelerator.
Piper is a CLI for local text-to-speech, with downloadable voices, WAV output and options for your own speech applications.
A skill and scripts for transcribing audio through Yandex SpeechKit: audio preparation, Object Storage upload and asynchronous v3 recognition
Local speech synthesis for agents: a command and a skill voice text, clone a voice from a recording and render scores to audio, with no cloud
yandex-stt: long recordings to text with SpeechKit
yandex-speechkit-talk-recognition
A Node.js command-line tool that turns long conversation recordings into text through Yandex SpeechKit asynchronous recognition
A skill with two scripts: text to an .ogg voice message and voice to text through Yandex SpeechKit, with ten voices and three emotions
README in short
English-only models have an `.en` suffix; multilingual variants do not. The model table lists sizes and memory requirements, so users can shortlist models for their computer and compare transcripts of the same passage.