NAME

transcribe-audio.pl - Transcribe an audio file

SYNOPSIS

perl transcribe-audio.pl [FILE]

DESCRIPTION

This script transcribes an audio file using the OpenAI API using the `whisper-1` model.

It's more or less equivalent to the following curl command:

curl https://api.openai.com/v1/audio/transcriptions \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: multipart/form-data" \
  -F file="@./examples/data/speech.mp3" \
  -F model="whisper-1"

In the examples/data directory, you will find a sample audio file called speech.mp3. This is the default file used by the script. The script should run in a second or two and print the following to the console:

The quick brown fox jumped over the lazy dog.

OpenAI cost for this is less than a penny.

Alternatively, you can try this with examples/data/englishminstershereford_1_grierson_64kb.mp3:

perl transcribe-audio.pl examples/data/englishminstershereford_1_grierson_64kb.mp3

The above file is 7.7M, roughly 2,700 w0rds, and transcribes in less than a minute. At this time, the cost to transcribe this text is about 10 cents. You might want to redirect the output to a file for later processing.

perl transcribe-audio.pl examples/data/englishminstershereford_1_grierson_64kb.mp3 > output.txt

SUPPORTED FORMATS

The OpenAI API supports the following audio formats:

• flac

• m4a

• mp3

• mp4

• mpeg

• mpga

• oga

• ogg

• wav

• webm