llm-groq-whisper

Transcribe audio using the Groq.com Whisper API

Installation

Install this plugin in the same environment as LLM.

llm install llm-groq-whisper

Usage

Run transcripts using:

llm groq-whisper audio.mp3

llm groq-whisper --help

Usage: llm groq-whisper [OPTIONS] AUDIO_FILE

  Run transcriptions or translations using the Groq Whisper API

  Usage:      llm groq-whisper audio.mp3 > output.txt     cat audio.mp3 | llm
  groq-whisper - > output.txt

  Examples:      # Basic transcription     llm groq-whisper audio.mp3

      # Translation to English     llm groq-whisper --translate audio.mp3

      # Transcription with specific model and language     llm groq-whisper
      --model whisper-large-v3 --language fr audio.mp3

      # Detailed JSON output with timestamps     llm groq-whisper --response-
      format verbose_json audio.mp3

Options:
  --key TEXT                      Groq API key to use
  --model [whisper-large-v3-turbo|distil-whisper-large-v3-en|whisper-large-v3]
                                  Whisper model to use
  --response-format [json|verbose_json|text]
                                  Response format
  --language TEXT                 Language code (e.g., 'en' for English). Only
                                  for whisper-large-v3-turbo and whisper-
                                  large-v3
  --temperature FLOAT             Temperature between 0 and 1
  --prompt TEXT                   Optional context or spelling guidance (max 224
                                  tokens)
  --translate                     Use translation endpoint instead of
                                  transcription
  --help                          Show this message and exit.

Development

To set up this plugin locally, first checkout the code. Then create a new virtual environment:

cd llm-groq-whisper
python -m venv venv
source venv/bin/activate

Now install the dependencies and test dependencies:

llm install -e '.[test]'

To run the tests:

python -m pytest