Skip to main content

A simple script that uses the Groq API to transcribe audio files using the Whisper model

Project description

Groq Whisper Transcribe

This script transcribes audio files using the Whisper model via the Groq API

Usage costs can be found at: https://console.groq.com/docs/model/whisper-large-v3

Usage

Install the Requirements

To install the necessary requirements, run the following command:

pip install groq-whisper-transcribe

FFmpeg

This tool requires FFmpeg for audio processing.

  • macOS / Windows — FFmpeg is bundled automatically via the static-ffmpeg package. No extra installation needed.
  • Linux — Install FFmpeg using your system package manager:
    # Debian / Ubuntu
    sudo apt install ffmpeg
    
    # Fedora
    sudo dnf install ffmpeg
    
    # Arch
    sudo pacman -S ffmpeg
    

On Linux, you can also pip install static-ffmpeg if you prefer not to use the system package.

Typical Usage

# Transcribe to SRT (English)
groq-whisper-transcribe audio.mp3
# Transcribe Spanish audio
groq-whisper-transcribe audio.aac -l es
# French audio to VTT
groq-whisper-transcribe audio.wav -f vtt -l fr
# Both SRT and VTT
groq-whisper-transcribe audio.wav -f srt vtt
# All three formats
groq-whisper-transcribe audio.wav -f srt vtt txt
# Plain text only
groq-whisper-transcribe audio.wav -f txt
# Custom output path
groq-whisper-transcribe audio.m4a -o result.srt
# Use turbo model
groq-whisper-transcribe audio.mp3 -m whisper-large-v3-turbo
# Set API key interactively
groq-whisper-transcribe --set-key
# Set API key via pipe
echo "gsk_..." | groq-whisper-transcribe --set-key

Arguments

Argument Description
input Input audio file path
-o --output Output file path (default: input with new extension). Ignored when multiple formats are requested
-f --format Output format(s): srt, vtt, txt (default: srt). Multiple formats can be specified
-l --language Language code, e.g. en, es, fr, de, ja (default: en)
-m --model Whisper model name (default: whisper-large-v3)
--set-key Set the GROQ_API_KEY in ~/.groq (accepts piped input or prompts interactively)

API Key

An API Key is required to use this script, and usage incurs a cost. This tool utilizes Groq because of its high speed and low cost. However, if you have the necessary hardware, it is recommended to use the official Whisper version, which is free

The easiest way to set up your API key is using the built-in command:

# Interactive prompt (key input is hidden)
groq-whisper-transcribe --set-key

# Or pipe the key directly
echo "gsk_..." | groq-whisper-transcribe --set-key

This saves the key to ~/.groq. You can also set it manually by creating the file or using the GROQ_API_KEY environment variable.

The file path for ~/.groq varies by platform:

  • Linux/home/<user>/.groq
  • macOS/Users/<user>/.groq
  • WindowsC:\Users\<user>\.groq

The file content should look like:

GROQ_API_KEY=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

Project details


Release history Release notifications | RSS feed

This version

1.0

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

groq_whisper_transcribe-1.0.tar.gz (7.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

groq_whisper_transcribe-1.0-py3-none-any.whl (8.2 kB view details)

Uploaded Python 3

File details

Details for the file groq_whisper_transcribe-1.0.tar.gz.

File metadata

  • Download URL: groq_whisper_transcribe-1.0.tar.gz
  • Upload date:
  • Size: 7.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.10.12

File hashes

Hashes for groq_whisper_transcribe-1.0.tar.gz
Algorithm Hash digest
SHA256 dd9bb5e548224d672a3640e3c3011128d4c7ecd51792080cf853d0d0607fc472
MD5 a180115dca9bb2f682c57a46edd1cd4a
BLAKE2b-256 373dad11f8a45713fe205360a6375118c650aff6776c187a7dd216f2306c9163

See more details on using hashes here.

File details

Details for the file groq_whisper_transcribe-1.0-py3-none-any.whl.

File metadata

File hashes

Hashes for groq_whisper_transcribe-1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 7674edc4216604a8e627daa20283cb245d0b04e4a089828080b344b82aa2a1f1
MD5 d23f7139a6781cff314755667e31d7d1
BLAKE2b-256 6a8eb1d7eae9b2e5df3f044ba3f3cd9266a1a6c6042ce69f0192bf8c8e25a17f

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page