Metadata-Version: 2.4
Name: cohere-transcribe-cli
Version: 0.1.0
Summary: Local CLI for Cohere Transcribe (open-source 2B Conformer ASR).
Project-URL: Homepage, https://github.com/twangodev/cohere-transcribe-cli
Project-URL: Issues, https://github.com/twangodev/cohere-transcribe-cli/issues
Project-URL: Model, https://huggingface.co/CohereLabs/cohere-transcribe-03-2026
Author-email: James Ding <pypi@twango.dev>
License-Expression: MIT
License-File: LICENSE
Keywords: asr,audio,cli,cohere,speech-recognition,transcribe,transformers
Classifier: Development Status :: 3 - Alpha
Classifier: Environment :: Console
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: End Users/Desktop
Classifier: Operating System :: MacOS
Classifier: Operating System :: POSIX :: Linux
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.12
Classifier: Topic :: Multimedia :: Sound/Audio :: Speech
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: ==3.12.*
Requires-Dist: accelerate>=1.13.0
Requires-Dist: av>=17.0.1
Requires-Dist: huggingface-hub>=1.13.0
Requires-Dist: librosa>=0.11.0
Requires-Dist: numpy>=2.4.4
Requires-Dist: protobuf>=7.34.1
Requires-Dist: rich>=15.0.0
Requires-Dist: sentencepiece>=0.2.1
Requires-Dist: soundfile>=0.13.1
Requires-Dist: torch>=2.11.0
Requires-Dist: transformers>=5.4
Requires-Dist: typer>=0.25.1
Description-Content-Type: text/markdown

# cohere-transcribe-cli

[![PyPI](https://img.shields.io/pypi/v/cohere-transcribe-cli.svg)](https://pypi.org/project/cohere-transcribe-cli/)
[![Python](https://img.shields.io/pypi/pyversions/cohere-transcribe-cli.svg)](https://pypi.org/project/cohere-transcribe-cli/)
[![License](https://img.shields.io/pypi/l/cohere-transcribe-cli.svg)](LICENSE)

Run [Cohere Transcribe](https://huggingface.co/CohereLabs/cohere-transcribe-03-2026)
locally on any audio or video file. 14 languages, GPU-aware, long-form audio
handled automatically.

## Install

```bash
pipx install cohere-transcribe-cli
```

First run downloads the 2B model (~4 GB) into the Hugging Face cache.

## Use

```bash
cohere talk.mp4                    # auto language=en, auto device, panel-free
cohere talk.m4a -l de -o talk.txt  # German, write transcript to file
cohere clip.wav -q > out.txt       # quiet: raw transcript on stdout only
```

Status (decode/load/inference time, real-time factor, chunk count) goes to
stderr. The transcript goes to stdout. `cohere --help` lists every flag.

## License

MIT
